[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Project to improve Debian support model



Katy Tolsen writes ("Project to improve Debian support model"):
> I've started a project with the aim to improve Debian's support
> model by designing software to help integrate and track all our
> support efforts and make a system that bridges the gap between our
> supporters, developers, and users in a synergistic way to make the
> most of our efforts and help us identify and solve issues more
> effectively and make sure the issues are documented and accessible
> to everyone in a way most suitable to their needs.
> 
> https://kathryntolsen.github.io/diss/

Thanks for coming to try to help improve Debian and for talking to us
about your plans.

This project may very well turn into something interesting, but right
now it seems to be mostly a combination of handwaving, and extensive
and ambitious high-level spec documents.

Do you have any prototypes / code / etc. that we could see ?

> For consideration of our developers, I ask for your input on this
> matter because I envision that this project can benefit development
> on the Debian system by allowing developers to make use of our
> proposed diagnostic tree model to supply such files with their
> packages that will help users file better bug reports and resolve
> issues with your packages.

reportbug already does some of this.  You should clearly reuse the
information already provided in packages, that reportbug already
consumes.  See /usr/share/doc/reportbug/README.developers.gz.

It's not very structured.  If you want more information in packages
(eg a more structured way to collect info), then you should propose
extensions to the reportbug per-package-information protocol.

You should propose only the minimum set of metadata that is needed
right now and expect to extend things later.

This goes for your project in general: you would be much better off
creating the Minimum Viable Product and then extending it later, than
by spending a lot of time writing documentation for how you expect the
distant future to look.

One concrete comment: I did see you mention here
  https://github.com/kathryntolsen/diss/wiki/DISS-Diagnostic-Trees
that you were planning to use XML as a metadata format.  Debian has
long disfavoured XML for these kinds of uses (for good technical
reasons).

I hope this is helpful and good luck.

Regards,
Ian.

-- 
Ian Jackson <ijackson@chiark.greenend.org.uk>   These opinions are my own.

If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.


Reply to: