Re: Creating a survey for improving the FSD

[Denis 'GNUtoo' Carikli]

On Wed, 28 Dec 2022 15:31:13 -0500
Craig Topham <craigt@fsf.org> wrote:
> What feedback do you have for the Directory itself?
I think it would be useful to know the limitations of the directory,
for instance what it doesn't do.

I already wrote some text about that that I published in this mailing
list. It explained that the directory doesn't audit binaries for
instance. It would be nice to have a place to put that text.

The second thing that would be interesting would be to have a place
that explains how to reuse the data if there are easy ways to do that.

For instance there are various tools to do queries on Wikidata, and
there are also python libraries to do that too. There is already
support for a Free software directory identifier in Wikidata[1] so if
we can extract some data from the directory, we might be able to build
tools for specific usages that combine information from both sources.
sources.

For instance it might be possible to add information in Wikidata to
match Guix packages[2] with free software directory entries[1] and
manage to flag out of date entries for instance. Though as usual
there needs to be enough interest in that for people to code
tools like that, contribute data to be able to do that, and maintain
all that data and code.

Guix also has very strict dependencies declaration, so we might be able
to extract that information with tools too, to see if some dependencies
were missed. Though here Guix's goals is different from the directory,
because Guix often has only one possible way of providing the
dependencies (a software might depend on bash or ash for instance
while Guix will choose one or the other).

Parabola has very interesting data on non-fsdg compliant software as it
blacklists certain packages from Arch Linux for various reasons[3] (not
necessarily FSDG issues, but why it's blacklisted can easily be parsed).

There is also some work in progress work by bill-auger to combine
data from this blacklist with the free software directory, probably
to check if the directory has problematic software[4].

Another possibility would be to use the free software directory data to
make FSDG compliant repositories, for instance by reusing existing
non-compliant repositories and filtering out what is not in the free
software directory. 

Though here the issue would be to make sure that the binaries would
really be free, as the directory doesn't take care of that. And such
repository would also give less guarantees that distribution that have
package definitions that can be used to modify and rebuild packages.
This could work for things like browsers addons for instance.

So basically making it easy to reuse the data could bring new
use cases and get more contributions in this way or increase the
quality of the data if people write tools for that.

References:
-----------
[1]https://www.wikidata.org/wiki/Property:P2537
[2]https://www.wikidata.org/wiki/Property:P6765
[3]https://git.parabola.nu/blacklist.git/tree/blacklist.txt
[4]https://git.parabola.nu/blacklist.git/commit/?h=development

Denis.

pgpOhZjhq3KCD.pgp
Description: OpenPGP digital signature