Re: SSPL or server side public license, GNU better update the AGPL

[Florian Weimer]

* Jean Louis:

> The AGPL does not enforce the publishing of the source code when used
> server side.

Why do you say that?

That company previously told their users that the AGPL would not
extend to the software the users wrote using the company's product.
Their software also lacked a built-in compliance mechanism.  They also
never had to figure out what the AGPL really means because they had an
asymmetric licensing arrangement with their contributors. 

Of course all this means that the AGPL was not very effective in their
case.  But that doesn't say much about other users of the AGPL.

Furthermore, I think their use of the AGPL was very much along the
lines of using the AGPL as the GPL variant designed to support an
open-core business model.  (The general idea is that the AGPL creates
so much legal uncertainty when applied to a particular piece of
software that commercial users are forced to buy a proprietary license
for the software to mitigate business risk.  Obviously this was not
the original intent of the AGPL at all.)  But that model broke down
when the company decided that their profitability was threatened by
those who had no problem publishing source code and didn't consider
full AGPL compliance particularly burdensome.  It did not help that
the company had said before that the AGPL would not extend to client
software, no matter what the client software does.

I do think that the AGPL turned out to be poorly worded and should
have used language similar to this GPL clause:

    d) If the work has interactive user interfaces, each must display
    Appropriate Legal Notices; however, if the Program has interactive
    interfaces that do not display Appropriate Legal Notices, your
    work need not make them do so.

That is, if their is a source code redistribution mechanism built into
the work, you must no remove it.  This would mean that anyone who
wants to use the AGPL for real has to implement the compliance
mechanism for it, and that automatically takes a way a lot of the
ambiguity.  It also makes it clear that the AGPL is not really
suitable for things like in-process database libraries or Scheme
implementations.

But given the other factors in the case you raised, I don't think the
AGPL ambiguity was of great relevance in the end.  My impression is
that the company never had increasing the amount of free software
available to the general public as one of their goals.  The AGPL was
just a marketing tool.  And it backfired when it turned out that the
AGPL does not discourage commercial use in the way they assumed it
would.