Hi kpcyrd,
The various projects that make up our community often use identical or
similar terms and expressions for interrelated concepts. As a result,
I try to standardise the terminology and tone when I write the
reports. This has the advantage that it makes them easier to read and
avoids confusion for people who are not as knowledgable about the
finer-grained differences as we do. The benefit of doing this is
extremely apparent when you compare the published result and my very
first draft (which never leaves my local machine as it's not useful to
anyone).
However, doing this runs the risk that specificity and accuracy can be
lost, and I believe that is precisely what happened here. By
collapsing some points, I was not attempting to minimise the work that
you or the rest of the rebuilderd team has done, I was only attempting
to make the report more effective as a whole.
I believe the most time-efficient way to address this would be for you
to edit the post directly and to clarify the issues you outlined in
your previous mail. Whilst the report has been published, many people
read these reports well after they have gone live. The various blog
aggregators, where most of the readership comes from, will update
automatically in turn.
Your remarks about the unusually short delay between my two emails are
duly noted. This was just a quirk of my personal calendar this week.
> We made quite a bit of progress in these releases that other distros are
> eventually able to benefit from too, Arch Linux is currently the only
> supported one because it has been the easiest to integrate and already
> had the required tooling that eg. Debian is still working on. It's
> somewhat difficult to understand how the different items in the report
> are prioritized and the overall report feels rather debian-centric. The
> change from `verified` to `reproducible` is unfortunate because we as a
> community have also used that term for `builds deterministically` which
> is a lot weaker than what we've actually achieved. Instead we've
> verified and generated diffoscopes for the "real" packages we actually
> distribute and have identified and addressed issues that tests.r-b.org
> was not able to detect.
>
> There's only one independent rebuilder that rebuilds a subset of the
> Arch packages, I assume there'd be more if rebuilderd had more exposure
> in the reproducible builds project.
>
> There've only been 26h between the two mails, I'm not part of the
> academic community and I'm only working on reproducible builds in my
> personal spare time after my day job, I didn't have the time to check my
> emails that day.
--
o
⬋ ⬊ Chris Lamb
o o reproducible-builds.org 💠
⬊ ⬋
o
_______________________________________________
Reproducible-builds mailing list
Reproducible-builds@alioth-lists.debian.net
https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/reproducible-builds
