Mailinglist Archive: opensuse-project (371 mails)

< Previous Next >
Re: [opensuse-project] Re: Does the project have a publicly accessible project management tool?
  • From: Roger Luedecke <roger.luedecke@xxxxxxxxx>
  • Date: Thu, 17 May 2012 16:22:37 -0700
  • Message-id: <1337296957.3384.26.camel@Homodevil.Machine>
On Thu, 2012-05-17 at 08:37 +0200, Pascal Bleser wrote:
On 2012-05-17 01:07:31 (+0000), Jim Henderson <hendersj@xxxxxxxxx> wrote:
On Wed, 16 May 2012 19:34:15 -0500, Bryen M Yunashko wrote:

I believe that's the same people Roger's been talking to. I gather
they're interested in helping out. But I think ultimately what could
make it move forward is if they simply help out by actually implementing
an openSUSE Vibe instance and co-administrating it with us, thus we get
some expertise and they get some free advertising for their product.

Or it becomes yet another administration annoyance like the
forums and many more of our tools where only a very (*very*) few
can work on it, be it in terms of features or maintenance.

The current situation is already ridiculous enough, let's make
sure that we do *not* add more to it.

Yep. And now you mention it, I may (or may not) have been involved in
getting Roger in touch with Tracy (the PM) at some point.

I believe the knowledge gap is one of the big problems here. Not many
people in FOSS have heard of Kablink and it would be beneficial for the
Kablink/Vibe guys to have a large public instance. They'll definitely
get something out of it (provided the community likes the tool once it's
implemented.)

Well I don't care at all about the "Kablink/Vibe guys", let's
make sure that if we add yet another tool, it is something, for
once, that is nice to use and, ideally, that can be supported by
more than one or two people on the planet.

Injecting a tool into the openSUSE community infrastructure
primarily because it's "from Novell", no thanks, we had that in
the past, we've seen how that worked out (iChain?).

I was up on the Kablink site a few days ago, and it seemed like there
wasn't much new up there. I know Vibe is going strong, though, and it's
possible some of the workflow stuff could come in handy for us at some
point (and that's only part of the paid product IIRC - it's one of the
differentiators).

I'll circle with Roger and see where that was at. Thanks!

To be honest, I believe... no, I am certain it is a mistake to
go down that route:

(tl;dr but I'm trying to be comprehensive)

1. The spechul tool
~~~~~~~~~~~~~~~~~~~
Kablink has absolutely no traction in FOSS from what I can see
(I might be wrong, but I believe I would have noticed) so
it will be yet another liability -- let's not have the iChain
debacle again, please.
How many people will have expertise with it ?
(let's say five at most)
How many people will be able to help out with issues, downtime,
etc... ?
(let's say two at most)
If you're counting on "upstream" Kablink devs: I don't want to
hurt their feelings but, realistically, how long before it's
dumped ?

2. Yet another tool?
~~~~~~~~~~~~~~~~~~~~
Really? Don't we already have enough tools that create confusion
(e.g. openFATE vs Bugzilla, forums vs connect) and aren't
integrated in any way except for SSO ?

Adding more tools will primarily be highly effective at one
thing: create more silos, and we have way too many of them
already (irc, MLs, forums, connect, wikis, bugzilla, openFATE,
OBS, software search, docs.o.o, retro, indico).
I'm not saying we need a single tool for everything (don't give
me a straw man here :)) but cmon... :)

There is definitely a high risk of putting an additional service
into place that will become something like connect (almost
completely useless in its current state IMHO, and I don't
believe that will change), or create a divide like openFATE vs
Bugzilla (almost the exact same kind of content (features/ideas
vs issues) but in two different tools with no interconnection
whatsoever).
Also, from what I can remember from having seen it briefly, it
has a lot of overlap with other tools that are already in place.

Hence more confusion, more splitting of content and people.

3. The ugly tool?
~~~~~~~~~~~~~~~~~
Highly subjective, and I've only seen Kablink twice, it was a
few months ago (but hasn't evolved much since then apparently)
and... I mean... really?
I get it that you and Roger do like the tool but do you really
believe it is going to get a lot of traction ?

Kablink didn't strike me as a project management / tracking kind
of tool, are you sure there aren't alternatives that are both
more common in the FOSS ecosystem (and hence better known and
hence with a high chance in terms of acceptance) and better
suited for the job (also to reduce overlap with existing tools)
?

I don't count "sharing files in folders" as "project
management"; actually, a feature like that is a massive issue as
people will put things in there instead of github.

Actually there is already such a tool in place, that is used by
the boosters team (I guess it still is nowadays ^^) and has been
used a bit by the board in the past, which is "retro", an open
source agile kind of ticket tracker:
http://retro.opensuse.org/projects/boosters-work/milestones/20/goals

Retro is not bad, not that widely used though (certainly not as
much as e.g. Redmine, Trac, Jira/Greenhopper) but lacks more
fine grained permission setup, team/group handling, etc... to be
used on a large scale by the whole project (at least from what I
can remember, was already suboptimal to set it up for the
boosters + for the board).

But AFAICS they're thinking of using Trello instead.


To end on a more positive note (after that long rant)...

Don't get me wrong: some sort of task and progress tracking tool
could be highly useful if used in a central manner by everyone
and all the teams (artwork, boosters, infrastructure, possibly
OBS, marketing, board (some things need privacy though),
possibly factory, etc...) in order to have
- *one* place/tool where to look for tasks, TODOs,
progress/state
- interconnection/linking between tasks that go over the
boundaries of a single team (e.g. marketing depending on some
artwork being done before they can move on)
- a tool that doesn't do a dozen other things, as that'll just
create even more spilling of information and, with it,
confusion

I don't want to spoil your motivation and am thankful as
everyone else for your efforts to push things forward, but I
fear that all the issues I listed above are not considered and,
frankly, I have doubts that Kablink is the right tool for the
job, to put it mildly :)

cheers
Wow, there are tools there I didn't ever hear of:). I think you missed a
nice close-sourced one I saw AppStream devs using. That weird scattering
mess is why I was pushing for Kablink in the first place. Trash some of
the old tools, integrate others into Kablink via extension. Granted,
your criticisms regarding Kablink are not invalid... but it is the only
software of its kind that I have experience with, and the willingness of
Novell to help with integrating and developing it to suit our usage is
quite an incentive.

--
To unsubscribe, e-mail: opensuse-project+unsubscribe@xxxxxxxxxxxx
To contact the owner, email: opensuse-project+owner@xxxxxxxxxxxx

< Previous Next >