View source for Intellectual Property, Copyright, and related rights

{{notes}}
{{warning|'''BIG RED TEXT HELLO: This is not legal advice, or necessarily correct. Do not make legal decisions based on this.'''}}

I'm just someone on the internet, trying honestly to figure this out, ideally for general consumption, but any of this is at best a introductory read before diving into the real choices/implications/legalese.



=IP's practical concepts=

Intellectual Property includes (but is not limited to):

<!--
patents protect a way of solving a problem

copyright protect a way to express yourself

trademarks protect a way to recognize a product
-->


'''Copyrights''' are exclusive rights on an '''original work/expression''', primarily the rights...
: to display in public
: to reproduce
: to sell, and 
: to make derivative works.

Copyright's '''intent''' is largely to stimulate and protect investment in a creative process, by rewarding that creator for some amount of time.

Violated copyright is often about profiting from someone else's work.

Violated copyright can be proven by showing whose work it is.




'''Trademarks''' give their owner exclusive rights to use a particular '''symbol or phrase''' to identify and distinguish a particular product or manufacturer.

Trademark's '''intent''' is arguably largely to ensure a product is what it says it is. Not protecting anyone, yet
: For consumers in the same context, this is often not being misled by imitations, to have a decent estimate of quality of a product.
: For companies this is largely about putting money into a brand to be recognized in that way.

{{comment|(America also uses [https://en.wikipedia.org/wiki/Service_mark Service Mark] to point out when you're naming a service rather than goods. Most other places do not make this distinction.)}}

Violated trademarks are often about misleading people about brands. This is, fairly literally, about trading on someone else's name.

Violated trademarks can be proven by showing people are confused in their purchases.




'''Patents''' give their owner exclusive rights to use a particular '''invention''', i.e. a new idea and its direct implementations. {{comment|(In theory this is about inventions regardless of precise implementation - though in some fields that distinction is vaguer than you'ld think)}}

Patent's '''intent'' is to stimulate the willingness to invest in research process, by rewarding the time and money invested in development by giving the inventor a share for a limited amount of time.

Violated patents are about profiting from someone else's costly research {{comment|(...well, that's the intent, not always the practice, ''or'' legal measure used in court)}}




'''Trade dress'''[http://en.wikipedia.org/wiki/Trade_dress] can cover some ground the last three do not. 

It deals with recognition of a product, or even its packaging.
Which can be very fuzzy, and arguably is protection of brand and marketing.

So in intent, it's arguably closest to trademarks.

It is e.g. meant to stop counterfeiters from making things that look almost indistinguishable except from e.g. a slightly different logo that lets them "well ''technically...''" their way out of it being a trademark violation.

Fake handbags do that, for example.
For another example, when tablets were still newish, Apple sued Samsung when their tablet started physically resembling the iPad a little too closely - which would not be a patent violation due to different tech, or trademark because they ''did'' clearly mark it with their own brand.


<!--
Names are not copyrightable. They ''may'' be trademarkable.
-->

=Trademarks=
<!--
"Do I need to trademark?"

As in, can you use something your unique mark without registering it? 
Absolutely. In certain fields, it's quite common to use unregistered trademarks.

However, this is largely a tradeoff between "enough of a brand to be recognized" and "not enough of a brand to be imitated".

If you expect to ever be imitated, registering a trademark is recommended.

Note that you still need to actively defend your trademark yourself - but with registration this should be a relatively easy process.

Note that 
* no one else needs to have a re



"How do I trademark something?"

In some areas, use{{verify}}. In others, explicit registration. 
Which is localized to an area, so if you're an international company, you may want do do a bunch of these.


"Do I need to trademark my logo, name, tagline?"




Something does not have to be original to be trademarked -- but you might not ''want'' it as a trademark if it is not uniquely recognizable, just because that doesn't do much for you.


Confusability practically matters within distinct markets. If the same name is used by different companies to sell computers, cleaning products, and a stage show, no one is going to be confused, and it won't be a trademark violation.

{{comment|(There are some further details to this, such as that things that are just descriptive can't be trademarked, and pre-emptively monopolizing market that doesn't exist yet is also not)}}


You can't trademark a generic factual description, 

You will have trouble enforcing trademarks on generic words or symbols, unless and until you're large enough.

You will also have trouble enforcing trademarks you have not maintained.
-->


=Patents=
{{stub}}

<!--
'''Patents''' give their owner exclusive rights to use a particular invention.

In theory, anything that that couldn't me made in a few afternoons of mucking about in your garage,
and which also didn't exist already.


It can be hard to explain even in theory what patents don't cover - it's a weird and gray area in patent law.

Not specific devices (because that wouldn't restrict the most minor variations),
not their workings (because that would restrict anything remotely close)
or even concepts behind those workings.


...but something inbetween, because patents were intended to spur invention.

It was an inscentive for the person, to make something that would benefit society.

The patent system was a pretty good idea and good implementation at the time, 
because it tended to be used for more complex, long development cycles,
and for things that would be a serious investment for anyone. 

Because why would you invest your years and fortune,
if it were fine for your neighbour to copy the result the next day?

It would be good for society but ruin you immediately, 
which would mean that inventing something intricate would be stupid,
and only the very rich could even think of doing this.


'''Criticism'''

The patent system also doesn't allow for any inbetween.

Which is part of why there are fundamental problems with patents.

Many of which always existed at the start, but over time grew from acceptable to problem.

Consider:
* how exactly to test what a patent does and doesn't cover is often a little fuzzy.
: e.g. leading to favour those who can afford just ''having '' a court case


* The more patents there are, the more they are likely to overlap 
* the faster patents get registered, the harder it is to actually check their content
: which all just means there is little content checking being done
: and even where it's done, high quality checking is infeasible, so rare.
:: To paraphrase and steal, you could get a patent for stroking a cat. And maybe so could your cat.

* how to test for simplicity/existance is fundamentally messy (and certainly requires experts)
: leading to "that's so obvious it didn't look like it'd be patented"
: leading to "let's just patent everything in case our neighbour does it first"
: The more complex and abstract, the harder any of this is ''anyway''

* this means there is value for applicants to abuse the system in one of various ways, e.g. 
: trying to cover things as widely as possible
: sneaking in the obvious
: sneaking in details that you recently read somewhere else (hoping it's hard for other people to prove you didn't ''also'' develop it?).


In fact, many nontrivial products made these days ''almost certainly'' infringe on one or more existing patents.


It has been estimated that probably ''almost all'' software probably does.

Not just theoretically - if you wanted to, you could probably break the entire industry, and anything that depends on computers. So, everything.
It's just that no one is interested in going ''quite'' that far because it's not profitable/feasible to do so (except in a few predatory cases, see e.g. patent trolling).


Plus, whether the patent you claim is being infringed on is itself valid is another matter - in many countries patents can be challenged, and will be declared void if they turn out to be nonsense (sort of a fix around having to do waterproof checking up front) -- but only after a lot of time and legal expenses. 

So cost/benefit is the thing that roughly balances this system.


But you wouldn't call it the grease of this system, because there are so many examples where this distorts.

A '''patent assertion''' (the claim that a patent has been infringed on) can put a lot of pressure on small businesses even if completely groundless, and the assertion is usually a lot cheaper and simpler to ''make'' than to ''receive'' (investment risks, sales, etc.).

Small companies are often left alone, not for altruistic reasons or that of economic development (sometimes the opposite).
Often it's just that  they haven't been noticed yet.
Or have been noticef but aren't a legal-financial threat while small.
Or have been noticed - by patent trolls who are calculating the optimal point where the company
is on the tipping edge of being marketable and profitable. The optimal point of getting the most money out of them (and probably strangle it oops sorry guys).


As such, small tech startups are required more often to have a '''patent defense''', which refers to having a strategy when an opportunists call, to avoid you from going under. 
It is expected more and more to to have patent defense as part of their business plan, to convince investors.


Copyleft licenses have started adding clauses. For example, some have a 'patent retaliation clause' that prohibits licensees from filing patent suits against each other (doing so voids their own claim to the license, or apparently even the patent).



http://www.patenthawk.com/defense.htm

http://en.wikipedia.org/wiki/Patent_troll


http://www.law.com/jsp/article.jsp?id=1202432897434





'''On prior art'''

https://www.iusmentis.com/patents/priorart/

https://www.tms.org/pubs/journals/JOM/matters/matters-9106.html





-->
<!--
TODO: look at this bleh of text too


The patent problem. Largely ignored, probably because it only affects us indirectly.


The problem is that the system now primarily promotes using it the wrong way. One of the better known examples is that Bill Gates recognized how broken the system is, but that Microsoft has explosively started copyrighting because if it doesn't, it is very predictable that it would be in trouble later.
You can't blame them for it either, when the mechanics of the system feeds its own abuse. It means the system is rather flawed.

Even universities, which publishing politics aside are generally of a sharing and building upon results, have started registinging patents in noticable and growing numbers. [http://onthecommons.org/?q=node/1168]



A change in policy has made it even more broken. These days, the check for prior art and obviousness is, particularly in the field of software, too much work, and arguably too fuzzy. Primarily to keep up with submissions, patents are essentially only checked for form. 

If there is a problem with it, you can fight it, and a patent will get nullified if it was indeed nonsense to begin with. 
But there are legal costs involves, which means larger companies can invest in patent bullying. But that's the wrong word, because the system makes that likely to happen - if they have more patents, or handle more ideas that might potentially be patentable or profitable if patented, they _have_ to be defensive about it or a real bully will eventually crap on them in a big way.


Of course, that doesn't mean you can't pressure companies big and small to not release something because it's risky until the whole thing is cleared, and that only companies that have money to spare for the legal proceedings do.



It often actually doesn't matter _too_ much, because with software it's usually one implementation of a trivial idea that can be replaced with another implementation to the same effect.


Still, general ideas rather than processes or designs are also allowed. Some dumber ones I've seen on a a poster in a TA lounge in a university near here. It lists over a dozen patents that most of the web violates. If their owners felt like it, they could sue half the web.




http://onthecommons.org/?q=node/1168









The incentive for development is a good argument.
There are some fields where no one would ever develop, simply because years of hard
work would be immediately usable by everyone else. In such cases, the patent system
stimulates development.

Monsanto's seed patents are probably one such case.


Yes, in some cases that means developers who plan cleverly can make more money, and it can
steer developers to relatively idle fields.

But. It has also been pointed out many times and for years that some of the implications of
patents that are positive to one company or person at a time can be easily negative for society.
In particular, the real-world value of the immediate, full, and long-term monopoly varies wildly,
but in some cases, and with the now-common practice in patent sniping, this can easily throttle
sometimes even halt any related development - or even ''use'' of a technology.

This in particular when a patent describes something very wide or central, implying a cost to
society in the inability to use or develop  for the next few dozen years

Monsanto's seed patent is probably also one such case.




I don't believe in abandoning the patent system, but why are we not talking about tweaking it?

For example, recognizing that incentive for development seems useful, it might make sense to
allow for profits up to a few multiples of the investment, and/or some threshold above the investment.
The developer benefits, the

-->

<!--
===The various related rules===


-->

=Copyright=
{{stub}}

==Basic copyright==


Copyright intends to protects the cost and effort put into original works, from being exploited by others not putting in that effort.

Copyright makes it worth it to interest nontrivial time to make something creative and original.



'''Copyright''' is what gives you rights on an item to
* display the copyrighted work in public
* reproduce
* sell
* make derivative works


Since you own these rights, you can choose to authorize others to do some or all of those things, as well as transfer them completely.



'''Copyright does not need to be registered anywhere.''' (anymore)

Since the [https://en.wikipedia.org/wiki/Berne_Convention Berne Convention], copyright automatically applies on anything tangible you create, as soon as something is painted, written, or whatnot. 
: {{comment|(...'tangibly' / 'fixed' is a historic thing. Digital forms are a little less clear-cut, though laws are slowly catching up.)}} 
: even when it applies anyway, there may still be a few upsides to registration within a country



Phrases like "All rights reserved" seems to originate from the earlier (1910) [https://en.wikipedia.org/wiki/Buenos_Aires_Convention Buenos Aires Convention], and seem intended to denote that copyright in one country extends to all that adhere to this treaty (which was mostly the Americas), but seems to have no legal meaning elsewhere {{comment|(and isn't necessary since Berne, since it doesn't add anything to what Berne covers by default{{verify}})}}.   

: It's still a practical way to signal you might care enough to make a legal point of it. 




'''Copyright goes to the author'''

If you create a work by yourself, you now have exclusive rights - meaning no one else can use it without your permission. 

...with a few footnotes. And exceptions, in particular '''work made for hire''', which often comes roughly in two forms:

* employees 
:: In practice, if something falls directly within the scope of a job you are hired for, your employer gets the copyright
:: The wording also means your employer can't steal your hobby projects, 
:: related people going into business for themselves, creating a new distinct thing in the same field, is often fine, but take care

* commissions 
:: commissioned work is the creator's and not the commissioner's, unless you explicitly sign the rights over
:: It is in the interest of both parties (and also fairly typical) to have a contract with specific mention of transfer 
::: for artists, to specify details that may be challenged later, including attributions of style and originals
::: for commissioners to get rights, limited or not, in the first place.   Useful if you e.g. want to show it to people (which is frequently also in the artists's interests).
<!--::: may also help clarify other IP details, like I own this character but you own this drawing of them-->


You can make agreements otherwise, but you will want that in writing.





'''A copyright notice is not necessary''' for copyright to apply.

...anymore, in most places.

For example, in the US it was required until 1989, and [https://en.wikipedia.org/wiki/Night_of_the_Living_Dead#Copyright_status_and_home_media Night of the Living Dead (from 1968)] forgot to do so, and was considered public domain because of it.

[https://en.wikipedia.org/wiki/Copyright_notice Since 1989] in the US, copyright is default on new works. 


It's common enough to see a copyright notice:
copyright symbol, date of creation, and copyright holder name ("all rights reserved" used to be required, now optional).


This is not required, but 
is an easy signal that you ''care'' about the copyright,
and may make it easier to prove later that you were always the author,
and make make it easier for people wanting to license the work to contact you, etc.

It may also avoid the {{search|innocent infringement defense}}, which amounts to saying "..but I didn't know it was protected", which was never a great defense, and in most cases won't stick legally, it's just that if there is a notice, they ''more clearly'' have no case.



'''Further notes''':
* Copyright applies to public use, not private use.
: roughly because there is very little to exploit privately.

* Granting these rights has to be done in writing - just saying it's fine is not enough

* Individual (or even full) rights can be granted on a case-to-case basis, but also on a blanket basis.
: in some areas it's common to draw up a contract at the same time you sell an item

* When copyright holders are individuals, they're often not not very picky about granting them, particularly for individual and non-commercial use.
: When holders figure something may be sold, then you often deal in conditions and money {{comment|(quite expectable when holders are are businesses)}}

* You could make any agreement that holds up legally. A work can be sold, and/or rights granted. For example, a work...
: ...can be sold with all rights (e.g. original works)
: ...can be sold / loaned with some rights restricted (consider music, DVDs, etc.)
: ...can be due to royalty fees when publicly displayed/reproduced (consider music)
: ...can have specific uses granted, e.g. allowing it to be put on shirts
: ...can have uses granted for a limited area time, e.g. reprint rights in america for two year, internet post rights for one year, or whatever. Also relevant for 'limited edition' things. Usually only worth doing on valuable IPs


* "royalty-free" does not mean "copyright free"
: Royalty-free is a copyright agreement that says "we won't charge you for using this publicly" - and often with specific conditions, e.g. type of use (where, how)
: the agreement itself 
:: ''may'' still be for money - e.g. there are paid-for royalty-free stock image sites, often resellers with their own pricing model
:: or not - there is gratis royalty-free music online [https://www.youtube.com/results?search_query=royalty+free+music] without charge (often from the artists themselves)


<!--
* owning the original painting and owning its copyright are different things
: which is relevant to working under commission

* identifiable characters are copyrighted (and an interesting case)
: this is why fan art is often technically copyright infringement.
:: Companies tend to consider it a neutral-to-positive effect as it can help something to become and stay popular and sellable, and most businesses (or people) don't want the bad PR of going after their own fans.
: When it has negative side-effects they may start to care.
:: and it really can for them, e.g. when companies are unfairly accused of ripping of their fans


* watermarks themselves are copyrighted {{verify}}

* The concept of a new collection (also e.g. collages) is copyrighted
: that is, the particular combination is
: when it includes copyrighted material you still can't do much with it (publicly) until you resolve those
-->

==Variants==

<!--
{{stub}}


A copyright-related contract is an agreement like any other, 
so you can be as creative and specific about it as you want. 


So, sometimes you want to release something for very-public consumption.

Which means it's nice to do it so that other people can use it without risking any of legal bother.

You can say the copyright holder, and say that copies/derivatives must cite the original author, 
but also have waived most most individual rights so beyond that can use it however they please,
You are saying "I made this, look at it. That said, have fun making something else of it."


What you can do ''includes'' putting that next copy under more regular copyright,
which in some cases is against your idea.

You can also say that derivations must also be public,
e.g. by saying "it's fine to make derivatives, but they must keep this same license".

'''Copyleft''' (informally) refers to agreements where some rights are not only given to everyone,
but ''stay'' that way due to a chain agreement that holds the next copier to the same agreement.


For example, code released under the GPL or LGPL license implies that all derivatives must keep that same license.
A company can use it, but they cannot publicly use derivatives without showing what those derivatives are.


There are various subtleties. For example, the GPL means the ''entire'' software product you adopt it in must also be GPL, which can be incompatible with prohibitive to adoption. 

You can instead choose to release under the LGPL, which roughly keeps it as a separate thing,
so it comes down to "you can ''use'' it without legal implications on the rest of your product,
but if you ''modify'' this library, those , tho it you have to give those modifications"


Or you find that the GPL technically only applies when people ''release'' the product based on the code,
and they're only using it themselves -- to run a web service as a product. 
Then you might go to the AGPL, which has a specific provision for that.



-->

==Some practicalities==
===International copyright===
International copyright 
for a ''large'' part does what you expect, 
but it is ''not'' a given, and copyright specifics ''do'' differ per country.

And due to how the agreements on copyright developed over time,
countries basically apply their own copyright law to foreign copyrighted material.


Legally, a large chunk of the details that matter internationally seem to come from:
* [https://en.wikipedia.org/wiki/Universal_Copyright_Convention Universal Copyright Convention] (used much less than...)
* [https://en.wikipedia.org/wiki/Berne_Convention Berne Convention] (more or less enforced via TRIPS)
* [https://en.wikipedia.org/wiki/TRIPS_Agreement Agreement on Trade-Related Aspects of Intellectual Property Rights] (TRIPS)

===Copyright length===
Copyright runs out.

When they do, the works pass into the public domain {{comment|(though note that the term Public Domain is a US-centric one)}}. 


Because the length is settled, there is usually a date before which almost everything is public domain (either because there was no copyright or because it has expired). 

For example ~1922 in the US.

Internationally, ''most'' books before 1945, and ''every'' book before 1787.


More recently, it depends on:
* when it was made, 
* what type of work it is {{comment|(different lengths apply)}}
* whether the copyright was extended {{comment|(if possible)}}
* possible other specifics
* possible legal changes {{comment|(e.g. in the US and other countries, copyright length has been repeatedly increased, from ~30 to ~100 years)}}

===Copyright assignment===

<!--

The way you can transfer ownership of the copyright.


'''In open source:'''

In most cases there is no copyright assignment.
You have copyright on your contributions, 
the open source licence just also says how it can be used.


This is why it's easy to use and can't really be retracted, 
and why it's usually very hard to change license - you need to ask all the copyright holders to do so.

Which is often somewhat by design - when all contributions are done via copyright assignment,
you can change license at will.

(And this has happened, e.g. what Oracle did with Solaris after it was acquired from Sun,
which is why it sorta died a bit. It's an interesting case study)




http://www.newmediarights.org/book/what_copyright_assignment
-->

===Copyright and the internet===
<!--
Temporarily elsewhere

-->

==Infringement==

If someone infringes, you can of course ask them to stop doing so. 


This can make sense if you think a person just liked the work, and/or you can often easily come to some agreement or other.


In case of more willful abuse, e.g. someone intentionally profiting from your work, the best idea is often to consult a copyright lawyer to weigh your options.

In practice suing will often cost more money than you'll get in return (and in some cases may get you bad publicity), 
though it is likely a useful way of stopping the infringing.



===Uncopyrightable things===
{{stub}}


Under law, there are a number of things that can't be copyrighted.
The rules and lists vary, so check your local details.


In the US: (most of this is copy-pasted from [https://copyright.gov copyright.gov])
* Titles, names, short phrases, and slogans; familiar, everyday symbols or designs; variations of typographic ornamentation, lettering, coloring; listings of ingredients or contents. {{comment|(some of this is [[trademark]] stuff instead)}}

* Works that have not been fixed in tangible form of expression (for example, choreographic works that have not been notated or recorded, or improvisational speeches or performances that have not been written or recorded)

* Works consisting entirely of information that is common property, containing no original authorship (for example: standard calendars, height and weight charts, tape measures, lists or tables taken from public documents or other common sources)

* Ideas, procedures, methods, systems, processes, concepts, principles, discoveries, or devices, as distinguished from a description, explanation, or illustration

Particularly in the first and last case, but also in general: note that patents and trademarks are different things and may still apply.

===On crediting===

Say you add ''"I do not own this. All rights belong to the creator, ''[name]'' and ''[link]''."'', and possibly legalese-sounding stuff about not profiting off this copy. 
That fixes it, right?


No.

Only asking the person for permission, and getting it, changes anything.


Copying and crediting has no legal effect - you're still infringing, you're just being a little nicer about it.

{{comment|(...nicer than e.g. [https://en.wikipedia.org/wiki/Copyright_infringement#%22Freebooting%22 freebooters], e.g. copying popular videos to their own channel to get ad revenue for it.
And ''even'' if you make no ad money from your copy, you are still taking away from ''their'' ad revenue)}}


The ethical standard (and ideally the legal one too) becomes whether it affects the creator's ability to sell the thing, and keep making things like it. 
After all, the idea was protecting creation.


What the actual consequences are depends on what the artist/owner wants.

If it's the odd shared image on social sites, they might decide it does as much good as harm to them, so making a fuss may not be worth it.

If it does have a real influence on their bottom line, then they are more likely to take action (including when a ''company'' is the infringer, which is a clearer case of being exploitative).


This is also part of why "I gave them exposure" is not a legal argument (it's also often not actually true at all, but that's something else).


It's also why removing copyrighted material later doesn't nullify the effect it had earlier.

===Fair use===

'''Fair use describes exceptions to copyright''', cases where a use would normally be infringement, but for specific reasons is not.

...also the area of many myths, and more than a few legal cases.



The most useful angle to focus is whether it serves a different purpose, for consumers of the thing.
In legalese, whether it is '''transformative'''. 

Transformativity is certainly not the only factor, yet it's the one you would pick in introductions like this, because it ''tends'' to imply the other relevant ones as well.


If copyright's main intent is to protect a creator's investment in a creative process, then '''if something is not transformative''', 
it will easily play in the same market that the copyright holder of the first work does.

If so, it will easily take some of that same audience and have them view and pay for that derivative ''instead'' of the original - at the cost of the person who created that market in the first place. 

<!--
(Heck, the first creation may have made create a little niche market for more of itself in the first place.
-->


The clearest case would be digital piracy/freebooting - plain copies of the thing.

Yet there are many other types of content piggybacking that you can find now.

Rewarding low-effort copies amounts to not rewarding original creations, 
so if we collectively value the creation of new things of quality,
there really do need to be rules about this, and copyright is currently the most viable way.


On the other end of the reasons ans spectrum there are uses in research, reviews, parodies, many of which are considered transformative, not  because of the category but because they typically don't serve the same purpose, so chances are much better that it won't affect the original creation's market.



<!--
In a little more detail, '''checking whether a claim of fair use is valid''' is based on 
* the purpose/character of the use (e.g. commercial / nonprofit educational purposes),
* the amount/substantiality of the portion used in relation to the copyrighted work as a whole
* the effect on the potential value / market for the work and/or its owner


There are various cases where this feels (and is) stupid,
remember that copyright is made to protect creators making things.


And there's many areas where this really seems overkill. 

Say, if you're taking a song and changing all the lyrics, that's not transformative no matter how creative you are, because you copied another part of someone else's creation (the music) and are playing in the same market. 

That's not the easiest example, because that's more about whether you use the recording of the music. Cover songs are fine.

Also, music has a history 


There is some vagueness when it comes to painting after copyrighted work such as photos

Medium matters.  

Pictures, being a single thing, are a little more all-or-nothing, but the same idea.



That said, people messing around on youtube, or copying a picture into their blog, ''are'' more likely to get away with what is infringement than large companies are, but legally, there is no difference. 

You can argue that ethically there should be a, but this is a minefield of "how big is a brand allowed to be, and how do you measure that?"



Yes, putting a photo that isn't yours on your website is copyright infringement.



There's a whole list of practical reasons why you probably get away with that in practice, like that 

: random informative images is not something you make money on, so why make an issue of it?

: Its owner may be more likely to make a point of something they invested time in.  And less if it's just a photo they took.

: memes are frequently transformative
:: but this can be fuzzier, see e.g. https://www.publicknowledge.org/blog/copyright-for-meme-makers/



There's also cost-benefit for companies in the form of fan work, e.g. fanfic, self-published work like doujinshi.
When there's little of it or it's not in a market you care for it's probably slightly good for you, and it's not worth shitting on your fanbase.

When it's large enough, or playing in the same market, then that's a much harder call.
-->



====Examples and myths====



<!--
'''For a more interesting example'''

There's an American cop during the BLM protests that played copyrighted music on his phone, 
even stating that that way any video would be copyright infringement.

Might it be flagged on youtube? Possibly (it's not very good quality).

Yet you don't watch BLM protests on youtube for the Taylor Swift. 
Anyone playing it does not intend to violate copyright, they would be there for the politics.

It is likely any judge would agree with that? There's a good chance.
-->


'''"You're making money, it can't be fair use" / "I'm not monetizing, so it's fair use"'''

Monetization has no direct bearing on whether the use is fair use.

In that you can make money while ''not'' drawing market/audience away from the original creator.

Sure, in practice it will be hard to  do this, to have no legal issues,
and do it this consistently enough to have it be a stable income.


Practically, money being made is certainly still relevant to 
: how quickly people will want to look at your case,
: and how thoroughly people will try to find other reasons, 
: and how interested they may be in a legal battle in the first place.




'''"If I add 'this is fair use', 'no infringement intended', and/or pointing out the real copyright owners, that makes it safe"'''

In general, this legally changes absolutely nothing.

Because none of that changes any of the just-mentioned points about audience and market and such.

Depending on the case, copyright owners ''may'' be somewhat less likely to sue you, but often for other reasons.




'''Educational uses, use in research, use in news reporting is fair use'''

Yes, but not because because of what that field is.

And instead because of what you do in these field is easily transormative.

Say 
: A video fragment in the news mostly mainly shows the context.
: If you're studying a thing in depth, you may even end up showing much of the work, but most of your content may still be that study.


At the same time, it's absolutely possible, within each of those fields, to violate copyright.

It doesn't happen nearly as quickly, but still very easy if you ''are'' trying.




'''"Using short fragments makes it fair use"'''

It's not the fact that they are short.

But shortness can imply other things that ''do'' matter.

Most uses that need only a short fragment are things that don't play in the same market.
For example, showing a scene from a movie and then saying "this is good example, now here's a discussion, which is the actual content here".

This is a good part of why uses in educational, research, and news reporting tend to be fine,
and also part of why critical review, response videos, video essays, and such ''usually'' are as well. 

It relates to why parodies ''are'' quickly fair use, because its nature is the criticism of the actual content,
often distancing it from the content itself. 
Though the gradations often involved means parodies are a topic of their own.
If you imitate a thing and stick 'parody' on the title, that may not hold up if it gets to court.


Sure, in practice people will probably check how of that content there is, and compare it to the amount of shown content.

Show one or two scenes, then talk about it for a dozen minutes? Probably fine.

Show a dozen minutes, then make a short remark? Assume no. (In part because you could watch what amounts to mostly the the original wprk that way.)



'''"If changed more than 10% it's fair use / not infringement"'''

There is no magic percentage, or timespan, or such.
While contents matter, use matters more.

Maybe just ask a random person on the street.
Do they say thing A is similar in purpose to thing B?
Then you're probably on dangerous ground.



'''"If I talk over it it's fine"'''

Can be, but it depends.

Say, youtube response videos will often talk over the thing they're reviewing.

Or maybe show it muted. 
Or maybe show a short piece looping continuously.
These things are intentional, because it's enough for ''their'' content, and makes it hard to consume the original content as ''originally'' intended. 

{{comment|(Specifically on youtube, where claims are fairly automated, another point is also that the way the content ID system is enforced is rather biased to punish creators, so even if they're in their right, creators want to avoid this.  These masking (flip it, make it half transparent, put some text over it) techniques aren't sure to work, but it really can't hurt.)}}



Say, if you're responding most of the time, that's fairly clearly transformative.

Yet if you show an entire movie, you'ld need to do ''serious'' amounts of yapping for it to not maybe amount to a public showing. 
And people probably wouldn't watch the result.

Case in point, MST3K licensed most of the movies (and presumably found a few with expired copyright?{{verify}}).  
Not because it wasn't transformative - you ''can'' argue that it often was, and ''maybe'' they would've ''usually'' been fine.
Yet they made a decision that the risk and cost of lawsuits probably doesn't outweigh getting that license.

(RiffTrax was an audio-only thing because that removed all worry, providing the same fun with only a little syncing for you to do)




'''"If I change the lyrics, it's fair use"'''

That depends.

Music is actually a weirdly complex case, and has recently gotten messier and more detailed over time.

In many places, the composition and recording receive separate protection, with their own logic to them.

Sampling is its own thing that can get you into trouble even on an otherwise original creation.<!--
Or people that licensed that original creation because they didn't yet realize it was an issue.-->

Covers are also a complex topic.


That said, play a silly version with different lyrics, particularly in criticism, and you may well be fine.

If you're just making it sorta-kinda-different but mainly clear reference, not so much.


And, while the risk is yours, you probably don't want probably-stuffy lawyers making the decision for you.

Case in point, Weird Al, probably the most prolific parody artist since the eighties, 
asks for permission (and apparently licences{{verify}}) all of his songs.

He may not need to, but it's less worry than the legal nonsense. 

Apparently he does have a dozen songs where, while the law would probably be on his side, 
he just won't release because the artist didn't approve. 
Presumably it's a mix of principle and 'not worth the potential legal bother'.






<!--
'''"Parody is fine"''''

[[Parody]]'s an interesting variation. 


In many places parody does get stronger protection,
but the way parody is judged is a fuzzy matter.

It is the critical nature that is protected,
just pointing and laughing probably isn't enough.

Whatever exactly you mean with parody - criticism, comedic or strange variant,
tropey, these tend to be different enough to be transformative. 

None of those are required, or necessarily enough, but each helps it be be something different.


One interesting case is Rick and Morty. 
Its origins lie in Doc and Mharti, which a crude short without much point,
and the characters are clearly Back To The Future reference (time travel, Delorean-looking car)
but it's not exactly a.

Rick and Morty contains a closer take on Doc/Marty's dynamic (It's all new to Morty and not to Rick - part of how and why the show works), but that's also where the similarities end.


-->


<!--
'''"This song sounds like that song"''''

Copyright in music is a complex thing, because it has changed over time.

In the times of early jazz, covering someone else was taken as a compliment,
and it helps that recording wasn't very common.

These days get into 
that sampling is typically fine,
but that a bassline and chord progresion shouldn't sound too similar, 
and that drum patterns tend to be too short to play as much in the judgment.

-->


<!--
'''"Music on livestreams is a problem"'''

For example, video games using music they licensed from its owners (usually just for a few parts) is fine at home, but when that same game is played on a live streaming service it amounts to ''you'' broadcasting that song, so unless you or the platform owns a license, that'd be copyright infringement.

Fair use doesn't really enter into it - it's use won't be transformed.


If you talk over it to the point people couldn't enjoy the song for the song then you'll probably be fine legally. 
Even if not necessarily safe from automated claims.


There's also cases where creators consider the in-game music they own licensed.
For example, Sims music will be claimed on youtube.

In these streaming days this is less common, because people may won't show others your game if they know this will be an issue.





Internet radio is a funny variation, because they tend to stream mostly copyrighted material.

Which is little different from radio broadcast before it, and the legalities are actually quite similar.

Both broadcast and internet radio tend to work of an "I'm allowed to play music because royalties are getting paid" structure.

-->


<!--

'''Related issues'''

You may run into legal issues completely separate from copyright,
like if you have trademarks or trade dress that may associate you with a brand that doesn't want to be associated with you. 

In some cases this may lead to cases like defamation - which is why movies tend to only show no brands - other than the ones that have paid for it. 

Not because it's necessarily an issue, but because you can avoid a bunch of potential later liability management, with fairly little work.


If you are specifically parodying it, then

-->

===copyright claims===


===spurious copyright claims===
<!--
Spurious copyright claims, are false claims of copyright, often one of:

* overreaching claims, e.g. rights beyond what the laws allow

* copyright claimed on something in the public domain

* copyright claimed on something copyrighted by someone else



In the public-domain case the claim is easier to check.

Contested claims can be much more complex. 
Because there is no registration or oversight on copyright, it may not be clear who made something.
One common strategy is simply to make your work publicly visible as soon as possible.

https://en.wikipedia.org/wiki/Copyfraud
-->

===Criticism===
<!--



-->

==Copyright free, Public domain, similar==
{{stub}}


'''Copyright free''', most literally, means works for which no copyright currently applies.

Which can happen from
* action - explicitly given up 
* inaction - expired, or whatnot.
* legal implication - such as that some governments saying that copyright will not apply for certain types of their output - see also [[public domain]], below.
:: this also roughly means that copyright ''cannot'' apply.


In many cases, copyright on exactly the same cannot be (re-)claimed later,
so a copyright free work can be used fairly freely.

That said, you could use it in a new work that is then copyrighted.
Not even preserving the original author's name is required,
though this is often considered good style.


Consider that IP other than copyright may still apply.
Consider e.g. publishing with trademarks that you do not own.



'''Public Domain''' is arguably a more specific ''idea'', yet in legal terms arguably less meaningful.

The wider concept describes cases where no exclusive intellectual property / copy rights apply, 
due to being inapplicable, or expired, forfeited, or waived, 
or sometimes just practically that no one ''could'' make a good legal claim to own it{{verify}}.


Yet the thing we call "Public Domain" is a US-specific concept, existing as something of an addendum to US copyright law.

The idea doesn't appear in every nation's laws, and where it does, it may not be legally equivalent.

As such, works that are PD in one country might have some restricted rights in another.
(In fact, the practicalities seem to vary a little even within the US.)



In practical terms, the boundary between "entered enter the public domain" and "became free of copyright" are not legally the same.

Consider for example:
* via an outdated (not renewed, or not renewable) copyright
: example: [http://onlinebooks.library.upenn.edu/webbin/gutbook/lookup?num=669 1913 version of Webster's dictionary]
: see also https://en.wikipedia.org/wiki/List_of_countries%27_copyright_lengths
: also note that some countries let you renew (extend) it a lot easier than others.

* by legal implication, such as by being released by the US government, and a number of other countries do this too
: example: the [https://www.cia.gov/library/publications/the-world-factbook/ CIA world factbook]

* by explicit release
:: has been questioned - see below
:: example: the [https://web.archive.org/web/20170930060409/http://icon.shef.ac.uk/Moby/ Moby lexical resources]



'''Being public domain is not a license.'''

Nor is it the lack of one, precisely,
in that PD also describes works where no one ''could'' make a good legal claim to own its copyrights / intellectual property rights{{verify}}.
Anymore, or sometimes at all.


When and where copyright exist on new works by default (that is: most places since the [https://en.wikipedia.org/wiki/Berne_Convention Berne convention]),
public domain is largely about copyright eventually timing out.



'''There is a real distinction between a lack of a license, public domain, and and a license that grants all rights to the public'''

Not declaring a license on a new work will, in many places, default to giving you copyright.

You can stay copyright holder but say all those copyrights are granted to everyone{{verify}}

To say you release it in the public domain (assuming for a moment that you can ) is saying you no longer have a claim to it.



'''Public domain may not mean free of all possible lawsuits'''

...the above also means that someone else can take something in the public domain, change it slightly,
call it a new copyrighted work, and then sue you for violation because the original is so similar. 


If that sounds like frivolous american litigous fuckery, yes, but also it's a valid legal concern.

Say, Getty Images has repeatedly gotten into hot water over issues like this,
taking images people have explicitly donated into the public domain,
requiring people to pay to use them
using very copyright-like terms and under threat of making a court case of it.

It's not that they charge for public domain images - that's legal, if potentially sleazy when they pretend they own it.

It's legal threats over something they don't really own (but if you bend the wording enough, can make a case of),
an argument that wouldn't hold up on court, but will scare enough people.


<!--
Say, if someone works or even just rewords a public domain song, they would be free to do so, but that song would be copyrighted.

That in itself is sensible enough, but how exactly the public domain song is then seen is unclear, 
given the wild redefining ride of music copyright in the last decade or two.

Music is not the simplest example, because the 2010s and 2020s have been on a ''wild'' ride of redefining how intellectual property works,
and by now it seems like a day job to even understand it.
-->



'''It is questionable whether you can actively release something into the public domain'''
: e.g. US PD law only mentions it as an implication of having been dead for a long time (and having no heirs to lengthen the copyright). {{verify}}

Sure, when they try to release it this way, the person who released it is unlikely to care about copyright later - but they ''might'' be within their right to do so. That said, there is probably legal claim to be retroactively be annoying about it

Still, if true, there is a difference and it is not the handiest thing to use as a license.{{verify}}





''Roughly'':

If you don't mention a license, then it depends a little on the place, but assume it's copyrighted to its author - see the [https://en.wikipedia.org/wiki/Berne_Convention Berne convention] (there are further details, though). 
Or employer, if that applies.


A license that grants all rights to the public ''may'' technically still be under an individual's name, 
but relieves all (mentioned) uses of any legal implications


Public domain means it is no longer anyone's ''and'' there are no legal implications to use.



https://law.stackexchange.com/questions/8732/how-can-i-release-a-work-into-the-public-domain





'''Public Domain is not an international concept'''

That is, the ''term'' is from the US, and some of the described details are also US-specific.


The ''idea'', and the execution of that idea, lives in many places, yet the name and its details may vary a bunch.


As such, saying "I release this into the public domain" may not do what you want it to do.

In the US, and/or elsewhere. 

It may not technically do anything.

If you want to say you license it to the public, this is different from you don't want a license,
in what ''other'' people can do with it.   Same idea, different technicalities.{{verify}}


https://medium.com/copyright-untangled/public-domain-why-it-is-not-that-simple-in-europe-1a049ce81499


'''It is not clear whether "release into the public domain" is a thing.'''{{verify}}

It has been questioned whether, even in the US, 
you can explicitly do this at all in a way that is actually legally does what you want it to.

In ''practical'' terms it may be similar enough, 
simply because if you technically still hold the copyright,
it's clear you won't go after people.

It is also not clear whether you would win a case if you said "no actually it's copyrighted" later{{verify}}


'''You ''may'' not be able to release into the public domain'''

Sure, if you say so, the intent is clear. 

If the author technically still owns the copyright, they're unlikely to enforce it still. 
I'm not sure a case has ever happened where someone has tried, and what happened.


https://law.stackexchange.com/questions/8732/how-can-i-release-a-work-into-the-public-domain



'''Things changed over time'''
* Berne convention was a big change for many countries

* China had a fifty year period without copyright, but they certainly have copyright now.

* In the US, copyright by default is a relatively recent thing
: copyright being there by default on any newly created work, without explicit copyright statement, is only a thing since around 1989. Before then, things were in the public domain by default{{verify}}. 
:: For example, the movie Night of the Living Dead (released in 1968) was considered public domain because it forgot to add a copyright indication[https://en.wikipedia.org/wiki/Night_of_the_Living_Dead#Copyright_status_and_home_media], and US copyright law required this at the time.




Misconceptions
* the idea that 'public domain' means "when it's out there, no longer copyright", e.g. if it's out on the internet, copyright protection no longer legally exists
:: copyright absolutely applies. You ''may' have some issues getting all copies taken down, particularly if you can't prove ownership.
:: but the internet is not a copyright bleaching machine


* Online, 'copyright free' is often confused with '[[royalty free]]' {{comment|(a lot of royalty free works are definitely copyrighted, with some specific terms)}}.



See also:
* http://en.wikipedia.org/wiki/Wikipedia:Public_domain_resources




-->



====CC0====

Given the above, saying something is public domain may be less useful than specifically giving a license with terms.

Even if those terms are a short bit of text, it is more explicit, 
and it's also a fallback license for places where public domain, by that name, doesn't even legally apply.

CC0 is one of several things that does that {{comment|(something similar goes for the [[WTFPL]], the [[unlicense]]{{verify}}, and some others)}}


See also:
* https://en.wikipedia.org/wiki/Creative_Commons_license#Zero_/_public_domain

* https://wiki.creativecommons.org/wiki/CC0_FAQ

* https://library.fvtc.edu/CreativeCommons/PublicDomain

====WTFPL====

The Do What The Fuck You Want To Public License is intended to be equivalent to releasing into public domain.


See also:
* https://en.wikipedia.org/wiki/WTFPL
* http://www.wtfpl.net/

====The Unlicense====

Another license that explicitly tries to be the equivalent of public domain.

It's three paragraphs, roughly saying 
: you are free to do these specific things copyright usually prevents you from doing
: we dedicate copyright to the public forevermore, to the degree local laws allow
: use this work at your own risk


See also
* https://en.wikipedia.org/wiki/Unlicense

* https://unlicense.org/

==Semi-sorted (copyright)==



===Database rights===
<!--

When you make a collection/compilation, that collection can easily have its own copyright, separate from the question whether copyright applies or even can apply to the things it contains (e.g. for images you probably would, for wordlists you would not - words aren't copyrightable).


This is distinct from most copyright in that the standard is somewhat lower: originality/creativity is not required,
it is the diligence/work put in.


It should be considered distinct idea from copyright law,
: in the sense that arrangement, selection, and presentation may be protected by copyright,
: and the same time that the database as a whole can be protected by database right.

In other places this is [https://en.wikipedia.org/wiki/Sui_generis sui generis] - specifically put into law,
e.g. The EU's [https://en.wikipedia.org/wiki/Database_Directive Directive 96/9/EC], or [https://www.legislation.gov.uk/uksi/1997/3032/made this UK regulation].


Database right exists does ''not'' exist in e.g. the US, and copyright does not ''necessarily'' cover it,
in that you can argue over whether it is an original creation or not.

The TRIPS agreement says that copyright extends to databases.


Note also that some different licenses specifically address database rights, e.g. CC does so.


https://en.wikipedia.org/wiki/Database_right

https://en.wikipedia.org/wiki/TRIPS_Agreement

https://en.wikipedia.org/wiki/WIPO_Copyright_Treaty


-->

===Poor man's copyright===

<!--

[https://en.wikipedia.org/wiki/Poor_man%27s_copyright Poor man's copyright] is basically adding a date via bank, solicitor, or post service,
so that you can later support the claim something has been in your possession since then.

This doesn't prove it was made by you, and US copyright law does not recognize this method.

However, it can still be useful to counter frivolous claims of ownership by other people.



On the web, crawlers and archive.org can serve the similar role of proving that something has existed, and in a place owned by someone specific. 
Again, not proof, but it does help.


-->

===Content ID===
<!--

Content ID is a google/youtube system, basically a music fingerprinting system for "register music here to automatically check other videos for matches" uses, so that copyright holders can get revenue when their music is used.

As copyright infrigement was rife before it, this is a good thing for creators (assuming creators=owners, but let's call that details for another day), and google had to do this because they were sued for a ''lot''.

The controversy of the system is due to the way it is implemented as an alert system, but used as a legal system.
: little to no human review
: automatic takedown, or ads for the claimant's revenue{{verify}}
This leads to large companies using it as a system of intimidation,
making money of allegation that turn out to be invalid,
but most of all, proper legal recourse involving having to risk your channel and/or having to go to court.

https://en.wikipedia.org/wiki/Content_ID_(system)

-->

==See also (copyright)==
* http://en.wikipedia.org/wiki/Copyright

* http://www.copyright.gov/circs/
* http://www.ipo.gov.uk/types/copy.htm

* http://copylaw.com/

* http://painting.about.com/od/copyrightforartistsfaq/Copyright_for_Artists_FAQ.htm

=(Chain) (software) license comparison=

Note: Most licenses here are software licenses. 


Most are also '''chain agreements''', where copying/using implicitly agrees to holding up the same terms, which can form a copy-and-agree chain.

This is different from basic copyright, in which you only deal with the copyright holder, which is often the single original creator (or distributor) of the content.



==Potentially confusing points==
===OSS, FOSS and FS===
{{stub}}


'''Free software''' refers to an idea that '' 'frees the recipient from any legal barriers which would inhibit their freedom to run, study, modify, or redistribute the software - both modified and unmodified.' '' {{comment|(source: [https://en.wikipedia.org/wiki/Free_software wikipedia])}}

This deals deals with '''liberty of ''use'' ''' in the wider sense of the term - also described as ''' 'free as in speech' ''', libre, freedom in a general sense - a 'no one gets to say I can't use this'.

Free software licenses often reiterate these "yours to do with as you please, and hand it to someone else to do the same" points.




'''Open Source Software''', OSS, deals with the '''availability of source code'''. 
* Open source implies that you can '''look''' at its source.

* This usually also implies you can '''use''' that code for personal use, 
: ...but not necessarily mean you can run it a business
: ...but not necessarily mean you can make money off it
: ...but not necessarily mean you can use it with your neighbours
: ...but not necessarily mean you can redistribute mopifications

Whether you do or don't have those rights depends on ''other'' parts of the license agreement.

Sure, a lot of open source license ''is'' free to use, modify, or free software in any other sense.

''But not necessarily'' - in the 2010 and 2020s more and more business models are playing with exactly this distinction.


...meaning that '''Free Open Source Software''', FOSS {{comment|(also sometimes FLOSS or F/OSS)}}, which was always a thing, has recently become a little more relevant.

It is the think you were thinking of if you're a idealist:
: You can get at the code, you can change it, and code can be easily shared back - and may even ''have'' to be.




Notes:
* People like [https://en.wikipedia.org/wiki/Richard_Stallman Stallman] have pointed out that there are some distinct freedoms {{comment|(can mostly be seen as levels)}} that free software aims to make free:
:: The freedom to run the program for any purpose.
:: The freedom to study how the program works, and change it to make it do what you wish
:: The freedom to redistribute copies so you can help your neighbor.
:: The freedom to improve the program, and release your improvements to the public, so that the community can benefit.

* ...in that view...
:: open source gives fewer guarantees, ''free'' software gives more of those
:: free software is more an implication of the license you put on it
:: open source is ''mechanically'' easier to do - you put the source code in a place where people can read it. But may have implications from license as well.


* The difference between free and open software can be subtle in its definitions, but ''can'' be pretty clear in its implications
:: so look at what it tells you you can't do, what you need to pay for to do

* we often say 'open source' meaning 'free open source'


* '''neither freeness nor openness refers to pricing''' to the end user  (''' 'free as in beer' ''', ''gratis''). 
: As long as the applicable licenses do not specifically restrict it, free and open source may well be sold.
: you practically won't get people to pay for much more than distribution costs, when the code is available and they can reasonably copy/use the thing themselves
: that said, there are sensible and profitable business models based on free and/or open source software
:: the most common is probably selling support services, or easing administration at scale, which some linux distributions have successfully done.

* It has become more common to do things like
** dual-license stuff that works like "you have most freedoms for personal use, but commercial use has more details".
** use the [https://en.wikipedia.org/wiki/Open-core_model open-core model], in which most of the workings are open source, but some fancy features are not.



See also:
* http://en.wikipedia.org/wiki/Open_source_license
* http://en.wikipedia.org/wiki/Free_software_license
* http://en.wikipedia.org/wiki/Permissive_free_software_license

===Source license, non-source licenses===
Licenses ''may'' be explicitly specific to certain content types.

Many licenses are source-code licenses,
because they are written with terminology meant for source code.
Which is necessary due to the way source gets combined.


Recently, [[copyleft]] licenses specific to text, metadata, video media, audio media, datasets, and more have appeared.

===Copyright and copyleft, 'Chain agreement', 'viral license'===
{{stub}}

Basic copyright applies implicitly and default, in most of the world  
: You make something, it's copyrighted, as implied by the laws that apply to everyone in a country.
: {{comment|(see Berne convention and related)}}



In contrast, ''copyright '''agreements''' '' are between people. 

These agreements usually alter and extend a specific right, and often apply just to the (two-ish) parties directly involved in making that agreement.




For example, businesses classically deal with intellectual property via basic copyright: 
: there is a single content owner, 
: often its creator or their employer,
and you can seek them out and sign some agreement, often a "you can use this copy"-style copyright agreement.



Another way of going about that is a '''chain agreement''', where due to the wording of the license
: if you take any version of the content,
: you use it under the terms (rights, obligations) it was released under,
: where those terms include clauses that mean the license preserves itself (partially or fully)

No matter what derived version you are looking at, the license says that you can take it, 
as long as you adhere to the text to the license.

The wording means it makes legal sense without a central owner.



'''Copyleft''' tends to refer to construction that says 
: if you alter and use it, you must show your work.
: ...and you must use the same license on that

This is often a chain license. 
Each license has its own take on freedom to use, study, copy, modify, and/or distribute,
but they tend to say most of those must stay.


The variation leads to a useful distinction between strong and weak copyleft:

'''Strong copyleft''' perpetuates all its restrictions.
: The GPL is a common example because it is easily explained: If you pick up something GPLed, you may only release it under the same terms - no more, no less. 
:: In other words, no matter how much that GPLed content evolves, it will stay GPLed.

'''Weak copyleft''' means licenses where there are situations where fewer restrictions apply. 
: They still perpetuate themselves and say 'show your work', but have fewer requirements (details varying with the license)
: Examples: LGPL, MPL



'''Permissive licenses''' are typically seen as distinct from copyleft.
: They do not require you to distribute changes, and so allow you to close the source
: typically the largest obligation is to mention where you took it from.
: Examples: MIT, BSD, Apache, and others, 




Examples: 
* library licenses are often a mix. 
: Take the LGPL, it basically has a "you can use it as a functional unit without restrictions, and the license doesn't affect code outside the library -- but if you change the library ''itself'' you have to give back" sort of attitude.


* GPL is strongly copylefted, a nature which can be seen as both a strong and a weak point.
: Strong point in that it's a very simple way to keep the code in the community, which is good for the code itself.
: Critics call it a '''viral license''' (or even toxic): 
:: once source becomes GPLed, it can't really become licensed under something else
:: existing GPLed content is very hard to multi-license
:: ...legally entirely possible, but it's usually prohibitively annoying to do - because it means you need to find all past contributors, and get them all to agree.
: GPL also puts a lot of force on code you want to use alongside - it basically has to become GPLed as well.
: This occasionally means that GPLed software ''cannot'' be adopted into certain projects, because it already had code that is either proprietary, or also commonly, under a different license that ''cannot'' combine with the GPL -- which is what the term [[#License_compatibility.2C_e.g._GPL_compatible|notes 'GPL compatible']] is about. 
: (This is one reason to consider using the LGPL if you can)


There is an argument that historically, GPL was a very useful thing for open source to become a thing in a largely-proprietary market - that these days the same sternness is not useful, and there are more business models where more permissive licenses make equal sense - and lead to more collaboration and cross-pollination.

This is a philosophical debate about what ''exactly'' it is that you want to flourish in the long run.

===False open source===

<!--

'False open source' (sometimes fauxpen source) refer to cases where creator call the license an open source license,
but from which they have effectively ''removed'' from the existing rights and principles we understand as open source[https://en.wikipedia.org/wiki/The_Open_Source_Definition#Definition].


The typical example is "[https://en.wikipedia.org/wiki/Source-available_software source-available]" licenses - lets you ''read'' the source, but does net you ''use'' it in any real way - and it creators may sue you if you do use it.

This does roughly nothing to improve the software commons - and may work against it.

Examples include 
some of GitLab's licenses,
some of Microsoft's Shared Source Initiative,
the SSPL (see below),
various licenses crafted for specific software.


There are more debatable examples. Take signal, which is mostly open source
: whatsapp has used it protocol for years now
: the Signal app is open source, and you can run your own, but if you ''call'' it signal, or if you make it interact with the signal app, you ''will'' be sued. 

''Practically,'' you won't manage 

So the only use of having this open source is security reviews. And even then you have to assume.




The Server Side Public License (SSPL), are both interesting case, because it's basically GPL3 
plus a section that basically says "if you use this, you must open all things required to run it as well".

It is effectively more aggressively viral, and Elastic calls it doubling down on open, but both MongoDB and Elastic are using it not for community or the commons, but to defend against cloud hosters trying to make a half-closed service of it.

this may be a very good business model for them,
but it does not fit the concept we currently understand as open source (even if Mongo and Elastic will wat to continue calling it that),
because it ''removes'' the right to use the software in specific ways.



You can argue about the AGPL being a mild case of the same, but it only really says "network use counts as distibution",
which tries to close the perceived loophole in the GPL of "I copy it and alter it and technically never give someone the application, I just use the application to run a service." 

This is arguably a solution more than a problem - but a real argument to have.


-->

==='Substantial portion'===
Various licenses state that they apply when the software/document are used in full, alternation, or use of a ''' 'substantial portion' '''.

This is inherently fuzzy. 

Sometimes intentionally - it allows people to copy habits and good structuring ideas,
while not allow them to copy-paste the entire project.


===Derivative work===
<!--
In the issue of whether something is a derivative or a separate work, the letter of the law falls a little short.

For example, the GPL stems from a time of monolithic executables, and its wording reflects that {{comment|(also one reason the later LGPL deals with libraries better)}}.


So use the spirit of the law, because it's what courts will do {{comment|(...''when'' they are technical enough to understand the issues.)}}


In the end, it seems to be largely a question of how entangled it really is.


Derivation
* modifying a source file            (yes)
* adding a source file               (yes/usually)
* subclassing                        (yes)

* functionally independent           (too vague)

* interacting with different service
* interacting with different service

* statically linked together         (apparently so - derivative of both)
* dynamically linked together        (can be a good indication of yes, not necessarily)
* running in shared address space    (can be a good indication of yes, not necessarily)
* plugins - depends on what they really do. In some cases it's a runtime case of linking together, in others it's purely interaction.
* embedding (largely independently)  (often not)


The linking and plugin part seems to be summarizable as "Not the point. If it would be derivative in source form, it is"


For example, if you interact with a different service over a socket, that's not derivation.
If you run two things that would be useless parts until they communicated via a socket, it likely is.
But if they are units upon themselves (e.g. a server and a client for it), they're separate.


Yes, to programmers that list looks arbitrary, and it is. 
But the line has to be drawn somewhere.


'''Derived data'''

Data for which the generation was assisted by some software is often your own,
and the license of the software you used only enters into the legalities if it has specific restrictions on such use (which you agreed to when/by installing/using).

''However'', content you generate that is based primarily or fairly directly on other data, 
the new data will easily implicitly fall under the license of that original data.

This is an area of some greyness, as there are judgments of how directly, how recognizably,
how distantly and such it is based on the data.

As a rule of thumb: if the originating resource(s) cannot be recognized in the resulting data, you're probably safe. If the new data or parts of it are recognizably similar on the 
previous data, you may well be obliged under its license.
-->

===Multi-licensing===
Multi-licensing refers to giving a choice of more than one license to use.

Multi-licensing {{comment|(often just two, i.e. dual-licensing)}} is not directly about license compatibility.
You can multi-license using various incompatible licenses, because it is a "choose one of the following" situation, and the next person only agrees with one at a time <!--(multi-licensing is a disjunction of the two)-->.

<!--
This assumes you are not being bound by something else while attempting to dual license.
-->


When one of the two is open source or free license and another isn't, this doesn't exactly make make things easier, and may mean the software takes two different paths over time.
However, there are some situations and business models in which this makes sense.

To take a real-world example, the Qt library can be used in two ways:
* under the paid-for QPL, which allows you to use Qt in proprietary programs without having to release anything, or alternatively 
* under the GPL

Free software projects use it under GPL terms, and the core of the Qt project thrives from that. In addition, you can choose to use it in proprietary code (without having to open anything up) and have to pay for that ability. {{comment|(And you ''could'' develop a variant of Qt, but you probably won't as you can only use it yourself, and there's value in keeping it compatible with newer developments of Qt)}}


In many situations, dual licensing has proven a bit of an anti-pattern, because it doesn't really give you or your customers anything, and it does still make your life harder.

===License compatibility===
{{stub}}

Licenses consist largely of restrictions.

When combining two different pieces of software into a derivative work, both licenses have something to say,
and these restrictions may be directly at odds. 


"Derivative work" is a whole another can of words: 
When is it two things entangled into a single derivative thing, and when is it just two things talking to each other?
This was clearer in the time of monolithic executables, and vaguer now that everything is much more modular.

But let's talk about the clearer cases of taking two parts and smashing them together into a useful single product that you want to release.



These licenses typically only have things to say when you ''distribute'' a combination - mainly because of the logic behind the copyright law this is all duct taped onto.

...so no, there are no legalities to whatever tinkering you do in your cellar as an end-user.
But you should still care about it if you ever want someone else to using your nice tinkered thing.



====Legal basics of combination====
{{stub}}


When you combine code with different licenses, both the licenses apply at the same time.

And you can't have a legally self-contradictory combination. 

So various combinations simply are not legally valid (to distribute).





Broadly speaking, the largest difference here is permissive versus copyleft licenses

'''Permissive licenses''' like MIT, BSD and Apache says author attribution lives on, and little else.
: It says little else. In particular, it means that these allow re-licensing of the derivative work.





'''Copyleft''' licenses (and pretty much everything else) have a "same license for all derivations" clause.

That means any legal contradictions you find cannot be resolved, and the combination is considered just... incompatible.


This happens quite easily. A good example is the 4-clause version of the otherwise-very-permissive BSD license, 
which says ''"Advertisements mentioning this software must acknowledge its origin/contributors"''.

And even though I'm sure the people who wrote the GPL like the idea, 
the GPL says "no more, no less than we required" and this is an additional restriction it does not mention,
so this is legally incompatible.


You will have minimal trouble absorbing code from MIT, 3-clause BSD, and Apache 
into other projects with those same licenses. Or any other, really.

And while you can take code from MIT, BSD, and Apache into GPL project,
taking GPL code into an MIT, BSD, and Apache project is only really possible if you now change the latter project's license to GPL.




====So what's the license on a combination?====
{{stub}}


<!--
That depends on the licenses in question.



Many have "substantial part" wording, meaning you don't have to think about legal blah when you just take a useful helper function, 
adopt a way of structuring a project, or such.


If you do use or alter a substantial part, 
then it's typically down to "Same license for derivative work" clauses.

Most have one, which means the result is effectively dual-licensed.



For example, you can easily include MIT code in a GPL project.
In that the GPL will now additionally apply to the MIT parts.
Which main means your README or About box now has some more text, but nothing else is required.


For exactly the same reason, an MIT project will not remain purely MIT if you include GPL code.
Technically, the combination becomes GPL on the next distribution.


And note this does nothing to the original copy of that MIT source.

This is mainly a thing in that changes to that source in this derivative project
that can't easily make it back to the original project. 

This is how things many things accidentally fork off and die.


And this is only an issue in that it is a point of principle to stronger copyleft:
You use something from the community ''and'' use it in a product? 
Then giving it back to this community is a thing -- and then only because one of an explicit "distribution must include source code" clause.


-->
====Can you combine with proprietary licenses?====

<!--
Again, up to the licenses. 


The GPL not so much, on principle, though it allows for system libraries,
because they don't have to be distributed along.


A more interesting example is that there is BSD code in windows, some in old tools, and it seems to be confirmed that part of the network stack borrowed from BSD's.
Which the BSD license allows foe.




You ''can'' get combinations to work indirectly, in that 
* if you can get all the copyright holders to agree, they can make a singular exception
: this is copyright law itself at work
: it's also impractical


* If all copyright holders of the other software agree to ''relicense'',  you may be able to combine
: unlikely in practice:
:: you must get all copyright holders to agree. If that's only you, easy. If that's dozens of people, that's too much of a practical pain.
:: If they believed in permissive licenses, it would be already.
-->


====Unsorted (compatibility)====

<!--


Gray areas:
* Reverse engineering. It's a gray area because it's a sliding scale from "hey, I like what that thing does, let's make a nice imitation" and "I'm taking all the parts verbatim"
: It's evern funnier when you have all the source code to take directly. 
: You usually don't want to.



* There is no violation in combining things as an end user, because you are not distributing it yet.
: This sounds like a potential loophole, in that you could replace "distributing the combination" with "telling someone else how to do it in" or even with "write a script that combines the parts on another end user's PC
: The first probably won't get you into trouble. The latter clearly breaks the ''intent'' of the distribution-related rules, so is rather unlikely to hold up in court.




License compatibility does '''not''' equate, as some people assume:
* '' 'usable as an near-equivalent drop-in replacement for ...' '', nor
* '' 'at least as free as ...' ''

What it ''does'' mean is potentially more complex, because it depends on what the licenses say.



Mixing permissive licenses (BSD, MIT) means essentially no rules,
beyond having to include both licenses.


Library licenses were a thing, which say something like "If you only link with this library as-is, then our license has no implications on your codebase. ''Separately'', if you alter this library in the process, you must release (only) those code alterations back."


Stronger copyleft licenses specifically perpetuate themselves, and their restrictions.

If 


via clauses like "this particular requirement must not become more restrictive / or change at all" which may legally make the combination legally incompatible.

* Incompatibility with the GPL comes from the GPL 
: If another license has the same self-perpetuating clause, 








'''So what's the license on the new thing?'''

Depends on how, and what.

What: 
* how close the parts are
: for example, even the GPL allows for proprietary system libraries, because they don't have to be distributed along
: Distinct-purpose systems that communicate via API or ABI 
: Distinct parts that could not work independently 



The combination must be fairly intimate. It doesn't matter 

* for example, the GPL restricts the license options of the combination.
: ...but even in its case, then combined works (where there is intimate combination, not merely collocation in the same folder). 



How: 
* If you merged on a very low level, making a new thing with substantial fragments from multiple other things, then all the copyright from the old parts still apply.

* if you link things together, or separate them via APIs/ABIs, then you can argue that these are just systems that happen to sit closeby, and treat each part separately

* 





If you truly make something new from the old part intertwine the parts, it's now multi-licensed.


Strong copyleft may also require that th the new work be lice



Practically it also matters whether 
* a piece of software can have a combination of licenses apply

* merging code from another project has implications



There can also be a difference between intertwining code from two differently-licensed codebases,
and linking together the result of essentially-independent compilations.



Some licenses specify that there may be no legal additions - often to avoid extra clauses circumventing the original intent. This is easily a problem area, because licenses compatible in idea and/but with added clauses about something ''else'' (say, patents) would also not be allowed.
This is in fact the case with the GPL; it cannot be combined with a good deal of licenses as that would violate the no-more-no-less term in the GPL. {{comment|(In some cases, licenses may be incompatible both ways)}}. 

This can be seen as a strength of GPL, as it means you can't weasel content out of GPL. You can also see it as a weakness for much the same reason, and the resulting inability to combine GPL with various other licenses.

While you can check whether any pair of licenses are compatible, and more than one of those may be relevant, the term most commonly seen in relation to license compatibility is 'GPL compatible', largely because GPL is one of the most common licenses and is a favourite among freedom nerds. Other licenses are often:
* small-scale licenses (e.g. made for specific projects, which you'll just have to ask its creators about). May quickly have restrictions that combine badly (usually get forgotten because they effectively isolate themselves)
* BSD-style, which are permissive and have much simpler implications
* Completely non-permissive (say, under basic copyright) which tends to be implicitly incompatible with most any license


Some people use (in)compatibility to refer to related concepts.
For example, in the context of Wikipedia policies about concent licensing, theword compatibility regularly refers to the answer of whether copying the data into wikipedia may happen.
This differs in that it is a specific question, about a one-directional relation where practically only the source's restrictions apply {{comment|(and/because the target restrictions are permissive, and a known constant)}}. 
This is a common question with multiple licenses involved, but not the only one.

Transfer is often not symmetric, so many things are not ''mutually'' compatible. You can, for example, use BSD-licensed content to GFDL resources (such as wikipedia), but GFDL content cannot be BSD (would imply forgetting the GFDL terms).



'''In more practical terms''', the case of incompatibility matters to coders and to projects in general.
It may imply two pieces of source code cannot legally be used in the same product at all, or may have to be kept separate entities even though that's inefficient, a pain, or whatnot.

Note that this applies only to code that you do not own. When you own the source, you can release it under as many ''alternative'' licenses as you wish. The next person in the chain will choose the license that suits them best, and (only) that will apply in that chain from there. This often isn't very practical, but it can be.
-->

<!--
Technically, you can probably choose multiple compatible licenses to all apply at once {{verify}}, though this doesn't necessarily have much value.
-->

<!--
GPL compatibility is an issue simply because GPL is a very common license (along with LGPL, BSD). 

Chances are that at some point, you'll want or have to combine something with code under GPL. If something bigger can't accept you, they'll look for something else. If you can't combine to become more interesting, your project may be abandoned for worse quality ones that won't stagnate, and can be probably be cleaned up over time.



GPL is arguably one of the larger compatibility problem among copyleft licenses, because of its 'strictly no more and no less' rule about other restrictive clauses. {comment|(That rule is there largely to avoid circumventing the GPL; it makes it one of the stronger chain agreements)}} 

Of course, the other side of the problem is that people seem to like creating their own license, which is just asking for incompatibilities. Choosing unusual licenses is likely to work against you.

At the same time, the cases of GPL incompatibility are often minor nagging points for its potential users. For example, the old-style 4-clause BSD license is incompatible with the GPL purely because of its advertisement clause, which most coders or users won't care much about in the first place. This is the reason for the new-style BSD license, which ''is'' GPL-compatible.



Further notes:

* Perhaps the easiest set of compatible copyleft licenses is GPL, LGPL, the MIT / X license, and 3-clause BSD. If you release source under other licenses, it is less likely to be adoptable and/or supportable.

* Note this list excludes many others, but perhaps most notably GFDL and CC (though those are content licenses rather than source licenses, it's still annoying that this is incompatible with interesting released under GPL or even LGPL)

* Neither GPL nor GFDL are compatible with Creative Commons (some or all?), (even though the idea and implied legalities are much the same with some CC licenses, e.g. Attribution Share-Alike) {{verify}})

* The GPL is not compatible with the GFDL {{verify}}

* Also incompatible with GPL:
** Apache License (That is, Apache License 2.0 is compatible with GPLv3 but incompatible with GPLv2)
** Common Public License 1.0
** MPL (Mozilla Public License)
** Sun Public License


* The LGPL causes less incompatibility trouble than the GPL (depending a little on the type of adoption)



http://www.dwheeler.com/essays/gpl-compatible.html
http://www.gnu.org/licenses/license-list.html#SoftwareLicenses
http://blog.milkingthegnu.org/2008/04/gpl-for-dummies.html

-->

===When do requirements apply?===
Many requirements apply at the point you release, (re)sell or re-license the software (in binary and/or source form{{verify}}). 


They generally do not apply to personal or hobby use unless that is explicitly stated (and there are legal limits to what may be limited) -- which, note, is also the case for copyrighted materials{{verify}}.

===Issues in combining software and licenses===
{{stub}}
<!--
Somewhat hairy.


If you interface code from two pieces of software, you usually incur the license of both, usually meaning you can only interface code with compatible licenses. 

However, software with incompatible licenses that hook up via a run-time plugin or some exposed API system is usually not in any trouble. Sometimes people will legally agree, at other times they may still bug you.


In some ways this makes sense; if something was designed to interface, it likely doesn't mind being used, and in most cases, the producing company will want people to (to some degree) as an extensible product is usually a touted feature, which could not even exist if you had to interface directly with proprietary code.


In other ways, it barely makes sense at all; compare interfaces from libraries (binary, and inherent), code APIs (serialized or binary), networked APIs (serialized or binary). 

From a technical sense, they differ only in how much the code has to do some conversion for interaction, and the system as a whole becomes slower.


There seems to be one license construction exception to this: Library licenses like LGPL.

-->

==Source code licenses==

===Some comparison of copyleft and permissive licenses===
{{stub}}
<!--

tl;dr:

Permissive licenses are great when you make something and just care about it getting used.

Strong copyleft is great when you do not want a company to run off with your hard work.

Weak copyleft is a nice tradeoff between "changes go to the public", and you think strong copyleft would prevent adoption.
-->

{| border="1" cellpadding="5" cellspacing="0" style="text-align: center;"

|- style="background:#eee" 
!  
! style="background:#eef" | GPL{{comment|v3}}
! style="background:#eef" | AGPL
! style="background:#f7eef7" | LGPL{{comment|v3}}
! style="background:#fef" | MPL{{comment|v2}}
! style="background:#ffe" | Apache{{comment|v2}}
! style="background:#ffe" | MIT
! style="background:#ffe" |(new) BSD
! style="background:#fee" |public domain / unlicense

<!--
|- 
| Must notify of changed files
| {{greencheck}} 
| 
| 
| 
| 
| 
| 
| 
-->
|- 
| can copy/distribute without getting in trouble<br/>(see also next two points)
| colspan="4" | under specific terms (more than mentioning)
| colspan="3" | under basic terms (mentioning)
| [[not necessarily]]
| 

|- 
| Change must be distributed as source
| {{greencheck}} 
| {{greencheck}}, also on network use
| {{greencheck}} 
| {{greencheck}} 
| {{redcross}}
| {{redcross}}
| {{redcross}}
| rowspan="6" | is not a license, so {{redcross}}

|- 
| distributed code must have same license
| {{greencheck}}
| {{greencheck}}
| only for the library
| only for specifically MPL'd files
| {{redcross}}
| {{redcross}}
| {{redcross}}


|- 
| Liability and warranty disclaimer
| {{greencheck}}
| {{greencheck}}
| {{greencheck}}
| {{greencheck}}
| {{greencheck}}
| {{greencheck}}
| {{greencheck}}

|- 
| Cannot use names to endorse
| 
| 
| 
| 
| {{greencheck}}
| 
| {{greencheck}}
| 


|- 
| Patent provisions
| some[https://fsfe.org/activities/gplv3/patents-and-gplv3.en.html]<br/>(absent before v3)
| some[https://fsfe.org/activities/gplv3/patents-and-gplv3.en.html]<br/>(absent before v3){{verify}}
| some{{verify}}
| some{{verify}}
| some{{verify}}
| {{redcross}}
| {{redcross}}

|- 
| Trademark provisions
| 
| 
| 
| 
| 
| 




|- 
| could be monetized
| colspan="12" | {{greencheck}}<br/>If you get people to agree to give you money, sure - e.g. for support. There is often nothing to limit nerdy people from getting a free copy elsewhere, though.<br/>
{{comment|(there are other licenses that can prevent using free copies, or make it annoyingly hard)}}

|- 
| Considers jurisdiction
| 
| 
| 
| 
| 
| 
| 
| 


|- style="background:#eee" 
!  
! style="background:#eef" | GPL{{comment|v3}}
! style="background:#eef" | AGPL
! style="background:#f7eef7" | LGPL{{comment|v3}}
! style="background:#fef" | MPL{{comment|v2}}
! style="background:#ffe" | Apache{{comment|v2}}
! style="background:#ffe" | MIT
! style="background:#ffe" |(new) BSD
! style="background:#fee" |public domain / unlicense


|}



====Explanation of summarizing one-liners====


'''"Must distribute modified source"'''

This comes from a mindset central to open source: 
: "if you improve it, the benefits are for all, and those benefits must not be restricted".
: Avoids a lot of double work and wheel reinvention.

For example, GPL, LGPL, AGPL, MPL does do this, 
whereas with BSD, X, Apache redistributing modifications is considered good style, but optional.



This usually applies when the result is distributed (released, and also often applies when internally distributed within an organisation). 
In that case, you can only demand the source when software with this restriction has been released, 
...and not before then, or when someone is using for personal use.

In practice there are variations of how this applies, which are sometimes important. Consider:

* '''Libraries:''' Consider for example the the difference between the GPL and LGPL. 
: While the GPL basically forces your entire project open, the LGPL makes libraries a separate thing:
:: if you simply ''use'' another's LGPL library it has no implications on your source around it 
:: if you ''change'' that library it applies only to that library.

* '''Services''' 
: some people, reading "on distribution of the software", got around having to provide source by providing the changed variant as an service instead of as downloadable software. 
: If you don't want that to happen, use something like the AGPL, which is mostly just GPL plus some 'on network use' wording.



'''"Any distribution &rarr; Same license" / "Distribution with this code included &rarr; Same license" / "Code that calls into this library &rarr; same license"'''

The "Must use same license" means the license says "no more, no less",
mostly so that derivations cannot be stripped of the terms you set.

This is a central idea of copyleft.


There is a real question in "does this 'disclose with same license' apply..."
* ...when you distribute software that happens to include said code in any way

* ...only if you change it -- while using it unchanged (relevant for libraries) lets you treat it roughly as a distinctly licensed thing
:: this is roughly the distinction between [[GPL]] and [[LGPL]] - you can use LGPL libraries without having to open-source the thing that uses it, while derivations from the library must use the same license

* ...even when you don't distribute it, but merely use it as a service?
:: the difference between the GPL and AGPL - the latter says 'network use is distribution'


Or, as the Mozilla Public License (MPL) solves it, its code may be included by closed/proprietary software, and only the MPL-licensed source must be distributed along{{verify}} (similar to LGPL, but per file rather than based on the distinction "library or not").{{verify}}





<!--
More practically, put, "Does it apply when you include our code unchanged? Only when you improve it? Always, even when you use and not distribute it?"



For example, say want to support an obscure sound file format, or image file format. Say you find a library for it. Say you are licensed with GPL, and it is not. 
Does that mean you can't use it?
-->



'''"Maintain Attribution and Disclaimer"'''

Well, just that.


This is true for almost all copyleft, as well as permissive licenses (for the latter it's ''most'' of what they do)

There are some exceptions, like the [[unlicense]]



TODO

It would be useful to also have
* patent details

* trademark details

* "Must state changes" {{verify}}

* no liability, no warranty (pretty much present in everything)



See also:
* http://oss-watch.ac.uk/apps/licdiff/

====On patents====
{{stub}}
<!--
Say you author software, and you apply an patented technique that you own.

If you release code with a patentend technique, and release them with a free/open software license, 
then local law probably ''implicitly'' grants rights to ''use'' the relevant patent, even if the licence does not say so.


Note: only by ''your'' creation and release; contributors ''cannot'' open up use of patents of yours by them implementing a technique you patented, generally even if you then edit it{{verify}})


If a license says nothing on patents, you tend to still grant others rights to use that patent.
Other licenses make this explicit.


This avoids the case where it's called open source only in name, in that you couldn't make or distribute any changes.

As far as I can tell, no (common) license tries to make patents a means to actually restrict people from treating your own source as open source.


'''Patent retaliation'''

Around the mid-2000s some companies had a lot of fun patenting the silliest things, 
including things that others had made, and went to war suing each other with them.

And open source code does not prevent itself from being patented later - it's simply a separate legal issue,
but a very practical threat to open source if companies can go "we patent your idea".

Yes, 'prior art' should prevent that, but practice has shown someone else can patent something,
even when your software is clear and demonstrable prior art. 
Some of these things would only get resolved by going to court, and even if you may win, companies can afford a heck of a lot more easily than most devs.


This is why licenses started adding provisions like "you may change or use, give you do not sue anyone for using or distributing their own variant",
which you can could patent retaliation clauses.



'''Infringing without knowing'''

It's not unusual for you to implement something for which a patent exists, but you simply do not know it. 

In fact, patents exist for software concepts now considered so fundamental (say, {{search|US3380029|sorting numbers}})
that you could probably find all software out there in violation if you wanted to. 

{{comment|(And that's before you get into more modern examples of 'yeah that existed already' ({{search|US7844915|one implementation of pinch and zoom}}, {{search|EP1964022|(visual to show) slide to unlock}}) or apparent suing fodder ({{search|US5960411|1-Click buying}} was used by amazon to sue others from having fast checkouts in general), {{search|USD670286S1|maybe you can use this to cover all laptops and all smartphones}} but was {{search|NetAirus Technologies vs. Apple|merely used to sue Apple}}). And then there's {{search|USD670286S1|devices with rounded corners}} patent, which is only slightly less stupid than {{search|US1052270A|patenting the wheel}}.  Yes, each is oversimplified but this kind of legal technicalities is exactly why this is impossible to oversee. And yes, some of these patents are so broad that they cannot be enforced, meaning it's not worth patenting, but companies seem to treat this as a "can't hurt to try, in case another company succeeds"}}

Bad actors could probably bring all software development to its knees - it's just that no one has any interest in doing so.



Patents are about what you accomplish, so this isn't about the exact code (implementing it yourself would change copyright details, but not patent details).

This is unrelated to whether it is open source or not. 
The more complex it is, the more likely this is. 


-->


====On trademarks====

===Copyleft===

While not an exact category, 
'''copyleft tends to mean "freer than copyright", and usually also "more restricted than permissive licenses''''


====GNU General Pulic License (GPL)====
Must:
* Include license
* Maintain Attribution and Disclaimer
* Disclose (distribute) modified source (and any source that uses)
* Same license for derivative work

Can:
* use commercially
* modify
* distribute

Cannot:
* sublicense


Upsides:
* "you improve it, you tell us" as a shared-cost continuous development cycle
* Improvements are necessarily fed back to the community (...on release or sale)

Arguables:
* Applicability to documents is sort of vague due to the code-specific wording. See GFDL instead.

Downsides:
* Implicitly incompatible with various others
: which makes it hard to combine in codebases where other licenses apply - any you may want to avoid GPL if you think that may apply in the future (see also next point)

* The tiniest use of GPLed source forces everything around it to be open  - potentially an entire project
: [[LGPL]] is clearer on the point of how far it reaches when your software is a library

* doesn't come into play when hosting things as a service
: [[AGPL]] addresses that.



See also:
* [http://en.wikipedia.org/wiki/GNU_General_Public_License GPL]


=====Version 2 and 3=====
Version 2 is still relatively current.

Version 3 updates 2 mostly in the point that it would extend to hardware (apparently, {{verify}}) and would have a provision there are no DRM-like systems (apparently, {{verify}}).

====Similar/related to GPL====
=====CeCILL=====
[http://en.wikipedia.org/wiki/CeCILL CeCILL] is a GPL-compatible adaptation of the GPL to the French legal system.

There are also two variants:
* CeCILL-B, which is rather like BSD, X11, MIT licenses
* CeCILL-C, which is rather like LGPL


=====IBM Public License (IPL)=====

A variation of the GPL that:
* moves liability from contributor to publisher/distributor. 
* terminates the license on patent disputes
Incompatible with (but generally very similar to) the GPL.



Related to, but not to be confused with IBM's Common Public License (CPL).

=====IBM's Common Public License (CPL)=====
* Source license
* Maintain Attribution and Disclaimer
* Must Distribute Modified Source
* Derivative work &rarr; Same license

* Avoids contributors from sneaking in code with patents and then charing for it.

* Future reviews could affect earlier software (also true for GPL, though{{verify}})

* See [http://www-128.ibm.com/developerworks/library/os-cplfaq.html]


=====Affero GPL (AGPL)=====
{{stub}}


Roughly speaking, Affero takes GNU GPL's '' "when you distribute" '' and makes it '' "when you distribute or use it as a service." ''



Affero GPL is based on the observation that under GNU GPL, if you make alterations, host the thing yourself, and use it to provide services or provide an API, that's not ''distributing'' per se, so you don't have to share your modifications.


In these times where we have stopped actually owning software, 
just temporarily pay for the right to use it, 
'distributing' is no longer a thing.

If sharing modifications that you use was roughly the point, then this amounts to a loophole, that you may want to close.


When this does not apply to how your software might be used, then it's a bit of a draconian measure, and may only make some things more bothersome. 
...in that technically, even if you get it to work locally for a hobby project you technically must distribute your modifications. (Not that anyone's going to bother you for that)

 
AGPLv1 is based on GPLv2 -- and incompatible with it (because of the extra clause). 

(AGPLv2 is AGPLv1 plus a "AGPLv1 or any later version" note, meant to allow shift to AGPLv3)

AGPLv3 is based on GPLv3, and these two are compatible. {{verify}}
<!--
Note this also means AGPLv3 has patent provisions
-->



See also:
* http://en.wikipedia.org/wiki/Affero_General_Public_License

=====Server Side Public License (SSPL)=====
{{stub}}

https://en.wikipedia.org/wiki/Server_Side_Public_License

===Permissive licenses===

====BSD licenses (a.k.a. Berkeley licenses)====

All vesions:

Must:
* Include license
* Maintain Attribution and Disclaimer
* <strike>Disclose (distribute) modified source (and any source that uses)</strike>
* Same license for derivative work - to sustain the attribution requirement.


Upsides:
* (for coders) you can use this code without strong implications
* (businesswise) this code can be used in proprietary programs you sell, requiring only attribution 
Downside:
* (program/communitywise) vampirism is easy - anyone can use without doing someting in return, although this only really happens outside the (F)OSS communities. 
:: The common examples are Microsoft's IP network stack being directly adapted from BSD's, and Apple's OS X which is strongly based in FreeBSD and mostly closed-source. No sharing, but you can argue both are positive results to end users.


For the case of binary distributions, these licenses will require copyright notices to be placed in the documentation.

<!--(Applicability to documents?{{verify}})-->


Note that there are different versions of this license, and that similar (ISC, MIT) and derived (OpenBSD, NetBSD) licenses are often functionally equivalent to one of them. (See also: [http://www.gnu.org//philosophy/bsd.html]<!--, []-->)


See also:
* http://en.wikipedia.org/wiki/BSD_license



=====4-clause, "Original BSD License", "BSD license with advertising clause"=====
The original BSD license had four clauses. NetBSD still uses it. The clauses roughly require that:
* Code (re)distributions must retain the copyright notice, the clauses, and the mentioned disclaimer
* Binary (re)distributions must retain the copyright notice, the clauses, and all disclaimers
* Advertisements mentioning this software must acknowledge its origin/contributors (what exactly to mention is stated in each specific license)
* Names of copyright holders nor contributors may not be used in endorsements of (this/derived) software without permission.


'''Not GPL compatible''' because of the clause about advertisements.

The advertisement clause was initially intended for the BSD operating system itself. 
Once used elsewhere as-is, this means licenses were floating about with arbitrary mentions to the University of California, and many licenses with programmers' own names, which causes larger projects (e.g. and merging such software) to require any ads to mention dozens of names.

=====3-clause, "Modified BSD License", "New BSD License"=====
The "3-clause", "Modified BSD license" removes the third clause from the 4-clause version, making it:
* Code (re)distributions must retain the copyright notice, the clauses, and the mentioned disclaimer
* Binary (re)distributions must retain the copyright notice, the clauses, and all disclaimers
* Names of copyright holders and contributors may not be used in endorsements of (this/derived) software without their permission.

'''GPL-compatible''' {{verify}}

=====2-clause, "Simplified BSD License"=====
FreeBSD and others use just the first two rules, i.e.:
* Code (re)distributions must retain the copyright notice, the clauses, and the mentioned disclaimer
* Binary (re)distributions must retain the copyright notice, the clauses, and all disclaimers

This is generally the assumed version when people say 'BSD license'. 


'''GPL-compatible'''

====MIT license / X11 license====
* Maintain Attribution and Disclaimer
* <strike>Must Distribute Modified Source</strike>
* "Same license for derivative work" (which implies only attribution) {{verify}} 


Since the MIT license does not prohibit the use of the copyright holder's name in promotion, it's basically equivalent to the 2-clause variation of the BSD license


Note: When people say 'MIT license', they usually mean the license originally drafted for X11, probably the most significant license to come out of MIT, though of course not the only one used within MIT. 


See also:
* http://en.wikipedia.org/wiki/MIT_License

====Apache licenses (ASL)====
{{stub}}

* Maintain Attribution, Disclaimer
* <strike>Must Distribute Modified Source</strike>
* <strike>"Same license for derivative work"</strike> {{verify}}
: apparently same license for unmodified, choice for modified?{{verify}}


The Apache Software License (ASL), usually just Apache License, is a permissive license similar to BSD and MIT style licenses, 
in that use is free of most restrictions and modifications may be kept secret and sold if one so chooses.


The main differences to BSD/MIT style are additions:

* ASL is more more verbose and legally explicit about its intent

* ASL2 mentions what happens when there are [[trademarks]] involved
: "This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor"

* it's also clearer about patents{{verify}}}

* ASL2 adds some more explicit protective detail related to patents
: "If You institute patent litigation against any entity ... then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed."


Versions:
* Apache License 1 was GPL compatible (by mostly just being a rewritten BSD license), but 

* Apache License 2 
** became incomptible with GPLv2 (and GPLv1) because of the patent details - but this ''may'' be preferable to you
** ...but seems compatible with GPLv3, in that Apache-licensed code can be used in GPL3 works. {{verify}}


See also:
* http://en.wikipedia.org/wiki/Apache_license
* http://www.apache.org/licenses/GPL-compatibility.html
<br/><br/>

====Academic Free License (AFL)====
* Source license
* Maintain Attribution and Disclaimer
* <strike>Must Distribute Modified Source</strike>

Meant as a clarifying succession to Apache, MIT, BSD, and UoI/NCSA.

However, AFL-1.2 and AFL-2.1 are '''not''' GPL compatible. AFL-3.0 ''may'' be GPL compatible.
<br/><br/>

====UoI/NCSA====
{{stub}}
Meant as a clarified version of the BSD/MIT licenses

Apparently GPL-compatible.

* http://en.wikipedia.org/wiki/University_of_Illinois/NCSA_Open_Source_License

====Python Software Foundation license====
{{stub}}

* Source license
* Maintain Attribution and Disclaimer
* <strike>Must Distribute Modified Source</strike>

BSD-style - doesn't require source to stay open, but license must stay intact.

GPL-compatible.


Note that an earlier version, the '''Python license''', was ''not'' GPL-compatible (because of some mention of the laws of the State of Virginia) - though the name 'Python license' can also be used to refer to the newer version.


See also:
* http://en.wikipedia.org/wiki/Python_Software_Foundation_License
* http://www.python.org/psf/license/
* http://www.python.org/download/releases/2.1/fsf/

===Weak copyleft===

In terms of the restrictions a license imposes, there is plenty of room between permissive licenses (which have almost no restrictions)
and strong copyleft (which tend to have a bunch)


Consider e.g. GPL says changes to the code means the entire program must be shared, and under the same license.

Weak copyleft like MPL2 basically says "share the part of the code you changed, under the same license"




====GNU Lesser/Library Public License (LGPL)====
* Fairly specific to source
* Maintain Attribution and Disclaimer
* Must Distribute Modified Source (but not code that only uses)


LGPL is something like "GPL when you change, BSD-like when you use."

You have to publicise alterations to the library, but generally this just improves the library quality (may sometimes create a flurry of use and improvement) and rarely forces you to expose something you didn't want to.

Upsides:
* Code can used in a (larger) project without having to open the entire project now or later
Downsides:
* Allows legal reverse engineering (but so do most others)
* LGPL-licensed software can be used in proprietary software. This is a purist view, noting this gives proprietary development an advantage over free software


In a semi-open world, LGPL (or one of the licenses like it) is a fairly good choice of both making your code useful for most everyone and making it low-implication for the most people. 

A critical, pro-FOSS-and-anti-proprietary stance would note that you can be giving a proprietary project an edge by doing this since you can ''both'' use the library easily. The FSF takes this stance and suggests you use the GPL.


See also:
* http://en.wikipedia.org/wiki/LGPL



====Mozilla Public License (MPL)====

We now mostly talk about MPL2, rather than older versions.

MPL2 lies somewhere between permissive and strong copyleft, so is frequently called 'weak copyleft'.


Where e.g. GPL says changes to the code means the entire program must be shared, and under the same license, weak copyleft

Weak copyleft licenses like the Mozilla Public License 2.0 also require users to disclose their changes to the source code, but requires sharing of a narrower set of code. If an author reworks any of the original files, they have to release those updates when distributing the code and license them under the MPL. 



https://en.wikipedia.org/wiki/Mozilla_Public_License


===Semi-sorted===

====Open Source License (OSL)====
* Source license
* Maintain Attribution and Disclaimer
* Must Distribute Modified Source

(to check)

Not GPL-compatible.

http://en.wikipedia.org/wiki/Open_Software_License

====Common Development and Distribution License (CDDL)====

From Sun (can be seen as Sun Public License version 2; SPL was basically a revision of the Mozilla Public License).

Allows proprietary software in open source whole, apparently based on the idea that as an OS you couldn't do without closed-source drivers.

GPL-incompatible.


====Artistic License====


====Python Software Foundation License ('Python License')====
* Looks BSD style
* http://www.python.org/psf/license/

<!--
=Licenses mostly specific to single projects=

Licenses that are specific (often only slightly so) to a specific application or project.

Many are only though are sometimes useful in lightly adapted form


====Nethack General Public License====

====X.Net License====

====Eiffel Forum License====

====Mozilla Public License (MPL)====

1.x was not GPL-compatible because of some specific requirements. 2.x is compatibile with most GNU licenses, though essentially via this specific dual license. {{verify}}


MPL1 also had a weird clause about having to resolve legal disputes in California.




====Qt Public License (QPL)====


====Python license====


====zlib/libpng license====


====Apache Software License====


====Jabber Open Source License (JOSL)====


====Netscale Public License (NPL)====


====Apple Public Source License (APSL)====


====Eclipse Public License====
BSD-style?

http://www.eclipse.org/legal/epl-v10.html


-->

==General / content / non-source licenses==

===Creative Commons (CC)===
{{stub}}
Creative Commons refers to a set of licenses that allow you to choose the gradation of permissiveness. 


The basic CC licenses are the possible combinations of the following requirements:
* All: Maintain Attribution (and Disclaimer)
* All: Name the original creator
* Option: Share Alike: derivative work must be shared with same license 
* Option: Non-Commercial: work may not be used for commercial purposes
* Option: No Derivative Works (only allowed to distribute unaltered copies of the work) OR derivative works allowed (but imply same license)

So there's a a bunch of combinations, like BY, BY-SA, BY-ND, BY-NC, BY-NC-SA, BY-NC-ND 


Note that this carries ''very'' varied meaning. Say, 
: BY anb BY-SA are basically 'do whatever and have fun, just mention who originally made it'

: ND means 'you cannot do anything with this'. 
:: '''In terms of the commons or creation, this is just as controlling as copyright''' - you're just allowed to post it for free exposure.


As such, '''CC without mention of what kind means almost nothing'''.




CC has since created additional licenses, including:
* Some specific cases, say, 
** attribution+noncommercial+noderivative for music sharing
** license that have rather specific allowances
*** such as allowing sampling of the music but not reproduction as a whole
** ...and more.

<!--
* Encapsulations (why?{{verify}}), such as:
** [http://creativecommons.org/licenses/publicdomain/ Public Domain adaptation] (US-only?)
** http://creativecommons.org/choose/cc-gpl
** http://creativecommons.org/choose/cc-lgpl
** http://creativecommons.org/licenses/BSD/
-->

See also:
* http://creativecommons.org/about/licenses/meet-the-licenses
* http://creativecommons.org/licenses/



===GNU Free Documentation License (FDL, GFDL)===
The FDL (a.k.a. GFDL) can be seen as a variant of the GPL that replaces some terms that apply mostly to source code, and makes them about text in general instead.


The upsides include copylefting needs, and the ability to use other FDL'd content.
FDL is used in various open content wikis, including Wikipedia.


The FDL also pays more attention to the more physical nature that documentation may take, considering books and such.

At the same time, one downside might be the burden introduced when you distribute more than 100 physical copies: you must include the FDL text and the source form for the content (digitally or physically). This is a problem for some formats, like pamphlets. 

<!--
One common alternative is a Creative Commons license.
-->


See also:
* http://en.wikipedia.org/wiki/GNU_Free_Documentation_License
* http://www.gnu.org/licenses/fdl-howto-opt.html
Comments:
* http://en.wikipedia.org/wiki/Wikipedia:What_the_GFDL_is_not
* http://wikitravel.org/en/Wikitravel:Why_Wikitravel_isn%27t_GFDL
* http://home.twcny.rr.com/nerode/neroden/fdl.html
* http://people.debian.org/~srivasta/Position_Statement.xhtml


===Gnu Simpler Free Documentation License (GSFDL)===
A proposed variant of the GFDL,
basically taking out the Cover Texts and Invariant Sections requirements for cases where that may be annoying.

Seems intended for manuals, textbooks and such, though could be used for any text.

See also:
* http://en.wikipedia.org/wiki/GNU_Simpler_Free_Documentation_License
* http://gplv3.fsf.org/sfdl-draft-2006-09-26.html


===FreeBSD Documentation License===
A permissive license. 

Rather like the BSD license, but for text.

See also:
* http://www.freebsd.org/copyright/freebsd-doc-license.html



===Do What The Fuck You Want To Public License (WTFPL)===
A no-nonsense variant on Public Domain {{comment|(...without age conditions, question whether you ''can'' release something into PD, and such)}}.

http://sam.zoy.org/wtfpl/



===Free&nbsp;art license (Licence Art Libre)===
{{stub}}

See also:
* http://artlibre.org/licence/lal/en/


===Apple's Common Documentation License===
{{stub}}



===Lesser General Public License For Linguistic Resources (LGPLLR)===
{{stub}}

Meant to keep data for (computational) linguistics open for use, and keep its derivatives similarly open.



===Design Science License (DSL)===
A license meant for sharing scientific data, images and such rather than text or source.



===Open Publication License (OPL)===
May be used in free or non-free form, and has some restrictive options.{{verify}}).

From the Open Content Project, which now seems to recommend CC.

See also:
* http://en.wikipedia.org/wiki/Open_Publication_License
* http://opencontent.org/openpub/


===Open Content License===
Somewhat restrictive, mostly monetary.

Arguably replaced by the Open Publication License.

See also:
* http://opencontent.org/opl.shtml


===Project Gutenberg License===
{{stub}}

Refers to the overall license applying to Project Gutenberg - which mostly notes that public domain laws differ between countries, and explains copyright details that may apply to some resources -- note that PG has roughly two types of resources: 
* public domain resources 
** which technically come with the non-public-domain Project Gutenberg trademark and license. You either abide by these, or strip any reference to PG from the text (for a fully public domain resource)
** some may have additional restrictions {{verify}}
* copyrighted resources that authors gave to the project (mostly for ease of distribution), which are usually under more specific terms.

The exact details of a resource are explained in a summarizing blurb that comes with the text. 

See also:
* http://www.gutenberg.org/wiki/Gutenberg:The_Project_Gutenberg_License

==Can I change code license?==
<!--

If it's someone else's code, this depends on whether the license allows it. 

With most licenses, that amounts to "no".




If you own all the code (e.g. because wrote all of it)
then there is no reason you can't release it under as many licenses as you can think of.

Well, it's not quite that simple. Particularly copyleft means that each new vesion
is a re-release in a chain. 
Those releases were released under existing terms and cannot be changed. 
If you keep them public, that's the license that has to be on it.
If you withdraw it, the license is still valid, you just made it harder to find.



But in general, if can get all the copyright holders to agree (a lot easier when it's just you), 
you can start releasing a new strain under a new license.

You could think of it (note: not legally accurate) as something you own under regular copyright 
ALONGSIDE you having released it under a specific license. 

...so yes, you can release it under another, 
and you can withdraw future releases  {{comment|(or release it under more than one - see [[multi-licensing]])}}.


This sort-of-inaccurate view also helps explain why this becomes complex if someone contributed patches:
you can see they contributed it under the license that applied at the time.

So if you want to change the license for the whole, you need to ask all of the contributors to allow the change on all the contributions.
This is technically possible, but practically often infeasible.



So all of my hobby projects on github I could easily change. 
Until I get kindly contributors.


-->

==See also==

Lists of licenses with comments:
* http://www.fsf.org/licensing/licenses/
* http://www.gnu.org/philosophy/license-list.html
* http://www.gnu.org/licenses/
* http://www.asiaosc.org/enwiki/page/Comparison_of_Open_Source_Licenses.html


Just lists:
* [http://opensource.org/licenses/ Open Source Initiative OSI - Licensing]



Other discussions, comparisons and comments:
* http://en.wikipedia.org/wiki/List_of_OSI_approved_software_licences
* http://en.wikipedia.org/wiki/List_of_FSF_approved_software_licences

* http://www.petefreitag.com/item/533.cfm
* http://slashdot.org/articles/01/08/16/131203.shtml
* http://www.eroj.org/linux/theory.htm
* http://www.softpanorama.org/Copyright/
* http://www-128.ibm.com/developerworks/opensource/library/os-license/
* http://www-128.ibm.com/developerworks/opensource/library/os-license2/

* http://www.croftsoft.com/library/tutorials/gplmpl/
* http://www.gnu.org/philosophy/netscape-npl.html

* http://www.croftsoft.com/library/tutorials/promiscuous/
* http://www.croftsoft.com/library/tutorials/openmarket/


Unsorted:
* https://www.gnu.org/licenses/license-list.en.html

* http://opensource.creative.com/pipermail/openal/2004-May/007560.html
* http://opendefinition.org/licenses

* https://tldrlegal.com/

[[Category:Legal stuff]]