Re: BPMLOD and string metadata

>
> Sure.
>
> I should mention, btw, that this is not a simple matter of open preference
> – the group should have a strong argument for using emails for technical
> discussions that takes into account the people outside the group who will
> interact with it. (Email is typically used for admin related topics, like
> this one, though.)
>

I think it's in fact a difference whether a discussion is more strategic or
whether it is specific to, say, a vocabulary under discussion or an
emerging standard. This discussion here is strategic, indeed. But for the
latter case, if it is maintained in GitHub (and most vocabularies from the
wider context of this mailing list are, e.g., OntoLex, lexinfo, NIF or the
Metashare ontology), *and* the contributors actually engage with it via
GitHub, GH issues make a lot of sense. Talking about OntoLex: There are two
vocabularies under active development (and several others under
discussion), OntoLex-Morph and OntoLex-FrAC, and indeed, we tried to push
the discussion into GitHub issues, partially, at least. (I'm actually on
your side, here.) But eventually, this wasn't regularly used. And email was
neither. Instead, the recent development mostly occurred during telcos and
contributions/drafts/samples prepared for these telcos. So, at least as far
as OntoLex-Morph and OntoLex-FrAC are concerned, the primary channel of
communication are minute documents and the specification itself. That might
be a very different situation from i18n, and, indeed, development is
relatively slow. But as I said, trying to move the discussions from
meetings and minutes into GitHub issues in OntoLex was not particularly
successful, so far (although GitHub issues are taken into consideration,
discussed in the telcos, and commented with a response in the issue itself
-- if they occur), and the OntoLex community has a lot of overlap with the
people involved with BPMLOD.

But I guess it also depends on the number of regular contributors. For
these both vocabularies the number of regular contributors has never been
more than a dozen people, and some of them coming from the humanities
(e.g., lexicography, philology), where working with GitHub is unusual even
in the first place, not to mention working with GitHub issues. Enforcing
such a shift has obvious benefits, but it may also lead to scaring away
potential contributors. (Just as a side-note: The development of
OntoLex-Morph was originally conducted via Google drive, only, so moving
the minutes archive to GitHub in Oct 2021 was already quite an improvement
-- but to this day, GitHub acts more like an archive for minutes and
samples -- and for the draft spec, of course -- but not so much as the
primary platform for discussions.)

Given the cons and pros, that should really be openly discussed, either in
the group as a whole or among the chairs, at least. There is a benefit in
transparency, but also the risk of losing parts of the user/contributor
base -- and developing best practices without having their potential users
involved is clearly a bad idea. We cannot really do multilingual linked
open data without having the language people on board.

Best,
Christian