[License-review] Request for approval of the updated W3C Software license

[Rigo Wenning]

Dear Mr. Piana

On 2024-04-20 12:50:34, Carlo Piana wrote:
> thank you for your submission. I am replying in my personal capacity of Open Source lawyer, without any of my statements and opinions meant or be construed as an official position of OSI in any manner.
> 
> I have in the past expressed some reservations as to how the license, in its previous incarnations, is drafted, but since it has been approved and it has not been substantially changed, I will defer to the previous evaluation and will not reopen the discussion.

Thank you. In fact, only the copyright holder in W3C changes because of 
organizational changes. The license itself does not change at all.

I nevertheless take the opportunity to respond to your very interesting 
questions.
> 
> On a general note, and building from the experience of retrospectively reviewing all the previously approved licenses, I note some points that I submit to you and to all future drafters of similar licenses.
> 
> This license seems conceived more as a one-off license that a contributor to a standard should prepare when a reference implementation is submitted. However, an Open Source license in my reading should be conceived to be used and easily implemented in all subsequent modifications and combinations. I will suggest here after a few possible doubts that a downstream developer could have.

W3C standards serve a large variety of stakeholders. The license 
submitted is a very old license that was crafted in 1998. The only 
adaptation so far was to also enable the application to standards work 
to allow for more liberty.

So in fact, the license just restricts the use (and abuse) of the the 
W3C name (inherited from the MIT license that served as a template). It 
is used in a large variety of contexts, for software produced by the 
W3C, for contributions to test cases, even for W3C Specifications.
> 
> - First bullet. "The full text of this NOTICE". What is the "notice" part of the license? The entirety?

Historically, the main target of the obligation to carry the Notice was 
to carry the disclaimer. The second target was the affiliation. So 
"NOTICE" in our reading means the 3 bullet points plus the disclaimer. 
But I note the difficulty. There is no chance I could just iterate here, 
but I may come back once a clarification has been pushed through the 
relevant committees. We know about the difficulties to produce the 
disclaimer in certain situations. Note though, that the short notice in 
the second bullet gives in indication.

> - Second bullet. Someone getting the software and wanting to reuse may be confused, not understanding what is the document you refer to, which is not defined.

The technique used is web (of course). A document on the W3C website 
makes explicit reference to that license to indicate that this license 
applies. This document may e.g. use other licenses in part. In that 
case, this is a reminder not to forget those other required notices. The 
relation between License and Document/Software is fixed in the URI used 
on the document that wants the W3C Software License to be applied to it.


> - Third bullet. It refers to a "W3C document", but why, since this is a software license? Maybe you will put a clarification note as to where you host the document and what it is meant to be. This clarification should be available irrespective of the availability of the web resources, including by OSI public repository of licenses. All licenses should be as much as possible self-contained and self-sufficient.

W3C has two licenses: A document license on certain recommendations and 
the software license. The naming was adapted when people claimed that 
one couldn't apply that license to documents. Because W3C applies the 
submitted license to software and documents alike. This is done, when 
W3C wants to produce forkable documents, or forkable tests, or forkable 
software.

> - Disclaimers. This does not raise particular concerns, but may I point out that the disclaimer only benefits copyright holders and not all who distribute, promote the software as is and could be left out in the cold? By contrast, the GPLv3 license provides "...OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS THE PROGRAM AS PERMITTED ABOVE".

Yes, this is inherited from the MIT License and was one of the 
conditions for W3C to be able to distribute software and documents that 
were produced as part of the Web development. Note that this was 
inherited from the X-Consortium that produced reference code and did not 
want to be liable for that reference code.

A developer is free to add their own additional disclaimer, if they want.
> 
> Again, this is not meant to raise any objection, or to imply criticism to the drafting, but as a general remark to any who would like to create a new Open Source license.

Thank you for the remarks. As a standardisation organisation, W3C is 
rather a special case with particular needs and practices. I noted down 
some of your suggestions in case I'm around  for future modifications.

Best regards

  --
  Rechtsanwalt Rigo Wenning
  ERCIM/W3C Legal counsel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 840 bytes
Desc: OpenPGP digital signature
URL: <http://lists.opensource.org/pipermail/license-review_lists.opensource.org/attachments/20240423/0ce2f358/attachment.asc>