Re: debtags revival

[Jonas Smedegaard <dr@jones.dk>]

Hi people interested in reviving Debtags (Bcc'ed this email)

Please subscribe to the mailinglist debian-metadata:
https://lists.debian.org/debian-metadata/

Please also let's use that for our future conversations - Reply-To: and
Mail-Followup-To: set accordingly.

That list has been dormant for 10 years, and the few ever posters to the
list are no longer active in Debian, so I see no harm in us "hijacking"
that list for largely related purpose.

Here is below some context from a previous conversation between some of
us (only my parts, out of respect for the other peers in that private
conversation) about my interest in Debtags:

Quoting Jonas Smedegaard (2024-08-13 22:26:44)
> A strong reason for me to use RDF is that it is a universal, declarative
> language: I can process it with Perl tools, and you can process it with
> Python tools - we need only care about the *data* format.
> 
> And you can use RDF for [your pet use] and I can use it for package
> selections¹ or social psychology² - we can use same data tools and
> exchange skills across fields, because the language is the same, for
> data-driven systems.
> 
> ¹ One of my use-cases for RDF-ized debtags is to change boxer to also
> use rdf, based on a (long-time dormant) draft work:
> https://salsa.debian.org/boxer-team/boxer-data/-/commits/rdf
> 
> ² The past semester I have explored use of ontology design to help
> understand scientific theories: https://purl.org/learn
> 
> 
> > > I will likeky use the prov ontology to tie tag to each package release.
> > > 
> > > I imagine first expressing debtags as SHACL shapes, like this
> > > (I am new to SHACL so this might be totally bogus):
[...]
> SHACL is not my answer to our question, only an intermediary.
> 
> You asked if I would track relationship between a debtag and a specific
> package release with a simple triplet.
> 
> Yes and no. I will track stated facts, fill in the gaps, and offer a
> richer dataset consisting of a mixture of declared and reasoned data.
> 
> What is stated now is debtags for packages (not package releases).
> That mostly works, translating to a simple "for any release of the
> package".  But not always, so I think of declaring that "for any release
> of the package" as a default translation rule (probably using SHACL but
> I am also very new to that - the concrete expression is not decided yet)
> and then add exceptions to that default rule whenever we discover
> packages where the default is inadequate.
[...]
> > > I am working on getting the Rust-based triplestore oxigraph stabilized
> > > in Debian.
[...]
> Oxigraph has now entered Debian testing.
> 
> Next on my agenda for that is to extend that source package to also
> provide binary package for its Python3 binding.
> 
> After that I will likely look into packaging a small library to allow
> using Oxigraph as storage backend for python-rdflib.
> 
> 
> > > Instead of trying to get a new mailinglist, I just stumbled upon
> > > https://lists.debian.org/debian-metadata/ which debtags fits into and
> > > currently lays dormant.  If you agree, I will ask on that list if ok
> > > that we "move in" there.  That also provides us a virtual package for
> > > filing bugreports against.

 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/
 * Sponsorship: https://ko-fi.com/drjones

 [x] quote me freely  [ ] ask before reusing  [ ] keep private