> We already have a very low barrier, as everyone can contribute by just
> creating a pull request or if you have an issue write an email.
That's of course another topic but mailinglist really feel
"old-fashioned" (beside its hard to find anything and you can not have
good references to them in issues or alike).
I also don't claim its extremely hard to contribute, just suggesting it
could be better and would make the Felix project look more open and get
a better experience for even small contributions.
> Allowing to create github issues does not lower this - and again, we
> have a very low number of contributions, even if these would be 5x of
> what we get now, it is still low.
But especially given that, I want to ask, what does JIRA buys us here?
Is there relally a need for a high sophisticated ticket workflow with
many management boards or teams that plan actual work? Or is it in fact
just some place where from time to time some text is dumped because we
require a ticket somewhere?
> What about the already existing open issues?
As mentioned, at platform we simply have closed them with a hint to
using github and asking the author to reopen it there if it is still
relevant. This is also a good opportunity to get rid of issues no one
actually is interested in.
> What about all documentation and links that point people to jira?
People will notice outdated links and then one simply can adjust them, a
first naive search on the github repo (where I usually start when
contributing to a project) leads me after some clicks to
https://felix.apache.org/documentation/community/contributing.html whats
seem the only reference to JIRA at all with a total of 7 matches so it
seems manageable.
> More importantly, how do we manage versions?
Github supports milestones if one really want to / need to manage them,
again given the low contribution frequency does one really need a high
sophisticated management? Is actually any more versions than the current
main line be managed/worked on?
In any case as I mentioned the maven team has already migrated, with
much more repositories involved and the have different active main lines
(at least 3.9.x and 4.x) so I'm quite confident the felix team will be
able to work it out without much problems.
> Just to be clear, I am a fan of github issues but in this case I fail
to see the benefit and I am worried that we make one thing easier while
complicating another.
For me having one place, one login is already a benefit. Especially for
low-contribution frequency the overhead to create yet another account,
register to yet another mailing list (to receive a lot of stuff I'm
probably not interested in because there is no way to control the
messages) is already a big win.
Not counting all the neat integration feature like automatic linking of
issues/commits to PRs, autoclosing, github actions integration or even
only to embeds some code snippets in an issue with screenshots just easy
as well.
Am 21.08.25 um 07:09 schrieb Carsten Ziegeler:
We already have a very low barrier, as everyone can contribute by just
creating a pull request or if you have an issue write an email. I agree
that creating a jira issue is a little bit more painful as you need an
account - but I would argue that 97% of people interested in OSGi have
that account by now.
Allowing to create github issues does not lower this - and again, we
have a very low number of contributions, even if these would be 5x of
what we get now, it is still low.
I do not think that it is as easy as just making jira read-only. What
about the already existing open issues? What about all documentation and
links that point people to jira?
More importantly, how do we manage versions?
It might not be that much work, but certainly it does not come for free.
Just to be clear, I am a fan of github issues but in this case I fail to
see the benefit and I am worried that we make one thing easier while
complicating another. And then we just wasted effort.
Regards
Carsten
On 8/20/2025 6:49 PM, Christoph Läubrich wrote:
I found that the lower the barrier, the higher likely someone likes to
contribute.
Also the "effort" would be just to enable Github Issues (and
discussions!) and make the JIRA project readonly.
We did this a while back with Bugzilla at Eclipse and it worked out
quite well, if needed one can simply link to the old discussion so no
need to really migrate all the old stuff.
Am 20.08.25 um 17:17 schrieb Carsten Ziegeler:
I do not really care - but I am wondering if the effort involved to
migrate is really worth it? Especially given the low value of
contributions we are facing.
Just creating (or asking for) a Jira issue when PRs are done without
it, seems to be way less work.
And I doubt that we get more contributions (being it issues or PRs)
if we move from Jira to github issues.
Regards
Carsten
