> So, what about ... keeping issues on Bugzilla for the moment ...
Mickael I know you are very serious on this topic but like to ask if we
should not even move to GH-issues here, and even take a step further
throwing away the wiki and probably deprecate mailing-list usage?
I think we'll get there eventually, and it's good to plan toward that. But I'd rather go incrementally.
- with the current release-when-neccesary-cycles we could benefit from a
more dynamic development process, with maven versioning people can
always use older releases if necessary and we can release fast and often.
I don't get what moving bugtracker or mailing-list would change to that compared to the current state.
- there is only a handful of "valuable" bugs in tycho, that is things we
are really working on, most of them are just laying dormant. We could
take the opportunity to ask the original author if its still relevant
and move those to GH, no feedback-> close with message to reopen at GH
if suitable
- as Bugzilla allows to link GH in the "see-also" we could use this
light-weight alternative to link related tickets
Yes, that's one possible approach
- maybe its just me but I have the feeling we do not need very much
project-management at tycho, and we can even try to use GH-Projects instead
Probably.
- The Wiki is mostly out-dated and hard to manage. The information there
is still valuable. Moving to GH-WIKI would allow us to have this part
editable by more users and as part of the usual development
I've become anti-wiki and would advise that instead of maintaining a wiki as a "side project" which can easily be out of the radar and out of date, we just put documentation in the source tree, near the code. That has several benefits over any form of wiki
* Doc is closer to source, so people looking for docs are likely to be become familiar with source and contribution workflows, and turn into code contributor
* Doc contributions are tracked as regular contributions, so people who contribute to doc get thanked on release by looking at history, and that's a good thing to do
* Code contributors would also feel more inclined in updating docs when they update code.
- GH supports a new feature "discussions" that might seve as a good
alternative to the mailinglist.
Why not... But maybe a bit later when we're done with other stuff?
Last but not least I think it would bee good to step-up and show all
this stuff can work and we are not bound that much to all the simrel,
backward compatibility stuff as platform itself.
I think removing all those different tools, UIs and brittled information
could really improve the overall development experience.
+1
