Hello Ekaitz,
On 2025-07-10 08:01, Ekaitz Zarraga wrote:
On 2025-07-09 23:34, Tomas Volf wrote:
The commit message's purpose isn't to enable people to avoid reading
the
diff; that's an impossible goal. It's to contextualize the diff.
Which, btw, most of the Guix commits DO NOT do.
I believe we should spend more time in the space between the commit
headline and the part when the files are listed, which in most Guix
commits is empty.
For me, context is very important. When I git-blame something and the
only data I get back is:
gnu: Fix thing.
* file (thing): Fix it.
It's as good as an empty commit message.
I guess with the new codeberg approach, the "Closes #XXX" might help
adding some context but I'd still prefer if people spent some time
explaining the intent and the approach of their changes.
I feel that the economics of transferring knowledge and information is
off in a lot of domains.
One area I feel is underserved is a meta representation of 'this is what
I think this thing is',
which from my perspective would be a graph which represents numerous
overlapping and contradictory viewpoints.
An example of this would be a social-bookmarking tool which would allow
people to provide notes and tags (ie folksonomies) to express their
knowledge of a link.
I adored for instance how the old Delicious service would be able to
provide different understandings and annotators as a consequence of a
link.
I mention this not to undermine the importance of strong writing styles
and the use of governance to encourage throughput.
In some ways a lot of Lispers have higher knowledge and the gaps are not
only the ambiguity from your example but people not (yet) mastering
sufficient subtleties.
When consulting historical material it would be nice if there was a
living ecosystem which provides better context.
(Im just throwing out these thoughts, Im not sure I want to drown this
topic in prescriptions)
...
I advocated against the changelog format, but I'm ok some
less-syntax-heavy version of it. Which I guess is what we are doing in
Guix somehow, but that has to be standardized because there is some
subjectivity about it. Some committers are very picky, others are not.
Ekaitz
Kind regards,
Jonathan
