[directory-discuss] directory purpose (was: s/w that requires a middleman..)

[Anonymous]

Svetlana Tkachenko said:

> If
> http://web.archive.org/web/20160825192530/http://gnuradio.org/releases/gnuradio/gnuradio-3.7.10.1.tar.gz
> works for you, I believe -- someone needs to confirm this -- you may
> link to it in the FSF directory instead of linking directly. Is it
> identical to
> https://github.com/gnuradio/gnuradio/archive/v3.7.10.1.tar.gz also?

Those files are not identical.  Hashes don't match.

> C) Software is free, documentation is non-free, website is non-free.
> Pass. We should include the software (but not the documentation) at FSF
> directory in some way.

"Pass" is ambiguous here.  What do you mean by "pass"?  Are you saying
it's included in directory.fsf.org?

This case "C" is why the "non-free documentation" anti-feature exists.

Indeed it's a good idea to not reference non-free documentation in the
directory.fsf.org record, to avoid FSF suggesting a non-free resource
to users.  But the reference to non-free documentation should be
disclosed /somewhere/, otherwise how do you defend the "non-free
documentation" flag?

Anti-features should transparently state the rationale for flagging.
For the rationale to stand up to scrutiny, it needs to identify the
artifact at issue.  Ideally the record gets a new field to hold the
flagging rationale, which need not be presented on the project page
but rather on the anti-features page for those who go for a closer
look.

> D) Software is non-free, documentation is non-free, website is non-free.
> (I think we all agree here) Fail.

It's unclear what you mean by "Fail".

If a software package (e.g. MS Word) is licensed as non-free and the
user has no expectation of liberty, then indeed it'd be pure clutter
to list it in directory.fsf.org.

But what about instances of "case D" where software claims to be free,
or is licensed as such, but it's actually non-free due to
noncompliance?  If you don't list the package in directory.fsf.org,
visitors will simply believe FSF has overlooked the package, and that
there is nothing wrong with the software.  When in fact such a package
is the most important package you can list, of course along with
anti-feature banners.  This is how the GNU user is best served.

Of course it's still disputed as to whether directory.fsf.org is meant
to serve package users.  Some people apparently believe it's there to
serve copyright holders, which reflects the direction Ian Kelling has
taken the anti-feature flagging.  IMO it should be a resource to
directly serve GNU users, and not the status quo: a tool to mislead
users for corporate gain.


--
Please note this was sent anonymously, so my address will be unusable.
List archives will be monitored.