[License-discuss] Pars pro toto: a fundamental(?) lack in (MIT licensed) (jquery) java-script packages?

[Richard Fontana]

On Thu, 2 Jan 2014 13:59:37 +0100
"Reincke, Karsten" <k.reincke at telekom.de> wrote:

> Question: Is it really necessary to add the MIT license to jquery for
> using this javascript library compliantly?

Note: IANYL, TINLA.

> So, we might say that Javascript libraries are
> distributed, namely in the form of source code.

Certainly distributed, though minified JavaScript is more
analogous to object code. 

> As an example, let us take the famous and very widely used jquery
> library ( https://jquery.com/ ) as. It is 'maintained' by an
> organization ( https://jquery.org/ ). It is licensed under the MIT
> license
> ( https://github.com/jquery/jquery/blob/master/MIT-LICENSE.txt ).
> Like all other instances of the MIT licenses with a customized
> copyright notice, it contains the requirement: 
> 
> "The above copyright notice and this permission notice shall be
> included in all copies or substantial portions of the Software."
> 
> If one downloads this package from the official download page
> ( https://jquery.com/download/ ), one unfortunately receives a
> package (a flat text file, even in its' compressed version) which
> does contain "the above copyright notice", but does not contain "this
> permission notice". Instead of this, it only offers a link to a
> license interpretation ( http://jquery.org/license/ ), being inserted
> into the first commenting lines. 

Which says, significantly, I think, for the issue here:

 "You are free to use any jQuery project in any other project (even
 commercial projects) as long as the copyright header is left intact."

> Based on these facts, we tend to conclude, that each user, who embeds
> jquery into his own pages, has manually to add "this permission
> notice" into that "copy" of jquery to which he links his pages - what
> probably means to add the complete license text. But - as far as we
> can see - this is 'not so often practiced' - perhaps because asking
> for that seems to be a kind of nitpicking.
> 
> Nevertheless: if we want to take the open source licenses really
> seriously, if we want to act compliantly and to fulfill the open
> source license conditions very thoroughly, then - as far as we can
> see for the moment - we have to add the license retroactively.
> Indeed, currently we are testing concrete solutions to do so
> ( http://opensource.telekom.net/oscad/fileadmin/js/jquery-1.5.2.js ).
> But from a viewpoint of a programmer, this is a suboptimal solution:
> Modifying an already tested component only for adding the license
> text is an unnecessary source of error.
> 
> Therefore, we want to ask: 
> 
> Are we right? Do we really have to add the MIT license to an MIT
> licensed package which does not contain this license? Or is there any
> way to distribute the library to our 3rd. parties in exact that form
> we received from jquery? 

I have considered this issue before. First, if you're really bothered by
the uncertainty here you can contact the jQuery Foundation or other
relevant authors -- it is possible they are not aware that some of us
think about this matter.

My view: by not including the full text of the MIT license themselves,
the jQuery authors are by implication giving everyone else permission
not to have to include it either. There are many issues of open source
license compliance analogous to this one which I interpret similarly.
If license foo literally requires X, but I the licensor myself do
something that implies a more liberal set of permissions than the
literal reading of license foo would indicate, it would be
preposterous for me to hold my licensees to a higher standard than I
myself exhibit. (Of course there may be reasons why you'd want to
'retroactively' include the license text anyway -- for example, this
might be convenient or helpful for your downstream customers. But I do
not think you are *required* to do this.) 

This assumes that the authors in question aren't in turn using other
people's MIT-licensed code as to which the full license text was
included.  

Any other interpretation seems unreasonable to me. It would defy common
sense: why would a project deliberately choosing one of the most
permissive licenses adopt a 'gotcha' sort of interpretation designed to
impose a burden greater than mere notice preservation, the true policy
reflected in the MIT license? The alternative interpretation would
further a policy of making open source license compliance more rather
than less difficult, which I do not think is in the interest of jQuery
specifically or open source generally.

I also assume that the non-inclusion of the full text of the MIT
license is done deliberately to save space, suggesting that licensees
are expected not to include the full text.

If the jQuery Foundation or the relevant authors clarified the matter
in favor of the stricter interpretation, I might change my opinion on
this.

 - RF