[linux-dvb] [RFC] Reviewed development procedures DVBMaintainer

Johannes Stezenbach js at linuxtv.org
Sat Mar 10 02:23:27 CET 2007

Hi all,

I've been thinking hard the last days what I could possibly say that
would make things better and not worse.

Mission Impossible :-(

I apologize in advance if I step on someone's toes.

But I'm not going to point fingers at anyone as I think this
isn't a single person's fault but more of a collective fuckup.

On Fri, Mar 09, 2007, Markus Rechberger wrote:
> On 3/9/07, Steven Toth <stoth at hauppauge.com> wrote:
> >
> >I suspect the other person gave the best advice they could at the time.
> >Maybe that advise is no longer true or valid. Either way, the person
> >offered the best advise I suspect. Maybe in the future you should seek
> >advise from multiple sources to validate any long term development 
> >approach.
> everyone is welcome to reply, but usually not too many participate here.

Well, generally I have the impression that there is a lot of
activity on linux-dvb. And just like you don't participate
in _every_ thread, but just the ones which you are interested in,
so is everyone else.

No one participates in a discussion if
- they don't get what it's all about
- or they don't care about it
- or they expect to be flamed or ignored

Or the other way 'round, if you're interested in feedback:
- explain the background so that _everyone_ can understand it,
  not just the "core developers";
  if you can post patches instead of a description what
  the patches would look like, please do
- show people why it is important _for them_ -- no one cares
  if it's only important for _you_
- be patient --  not everyone has time to fully dig into it
  right away, so the initial comments are likely to be
  superficial and/or just plain wrong;
  usually if you can show someone that their comment was
  wrong, they feel challenged to dig deeper into it to
  give you better comments;
  you must be careful not to hurt anyone's feelings, otherwise
  discussion stops pretty quickly

Sounds like hard work? You bet it is!

> >Likewise, the other developers have an obligation to review your ideas
> >and discuss/describe the areas of concern, offering constructive
> >feedback. It's unacceptable to refuse a colleagues patch and not offer
> >any advise. I know this has NOT happened. I know alternatives were given
> >to you.
> >I suggest you stop the bickering and begin working with your Linux peers
> >again, help us to find the right solution.
> sorry I'm getting pissed, for every approach I presented for now I had
> some code as well and it gets nacked without improvement suggestions.
> Also it has to be said that other hybrid implementations aren't well
> implemented and still need some improvement..

About one year ago there was exactly the same thing: Clash of
opinions about how to integrate xc3028, and then a huge
(and to me completely incomprehensible) fight followed.

Apparently everyone involved was so fed up with it that
they didn't want to touch the issue for a year, and now:
- the issue is still unsolved (surprise!)
- last year's fight still hangs over it like a dark cloud
  and prevents any useful discussion

I have no clue about the technical issues of this proposal
and can't comment on it, however I am convinced that the
technical issues are minor compared to the problems in the
social interaction between the developers.

I've read through the irc log, and I couldn't help noticing:


- Both mrec and mkrufky seemed to have predetermined expectations
  on the outcome of the discussion:
  - mrec: they'll NAK it anyway
  - mkrufky: he won't listen to anyone anyway
- It seems mws made a reasonable effort to understand mrec's
  proposal, but most of the time mws and mrec are talking
  past each other, and are finally arriving at the conclusion that
  the other one is just too stupid to get it.
  Maybe a hint that irc is not an adequate medium
  for this kind of discussions?
- mkrufky coolly watches the action, apparently satisfied to
  see his preconceptions confirmed; otherwise doesn't seem
  to want to help the issue at all


- Apparently irc sucks. I believe the mailing list is a much
  better medium for this kind of discussions.
- The proposal posting sucked (according to my criteria from above).
  Especially since Markus says on irc he has working code,
  I can't understand why he didn't just post the patches.
  Then there would have been no guesswork, everyone could've
  seen immediately how it works and what the consequences of
  this change are.
- I cannot see any malicious behaviour in this. Sure Mike
  was pretty unhelpful, which I think was lame measured
  by the role he plays among the DVB developers. But in
  the light of last years xc3028 trouble I can understand
  that he doesn't have any incentive to help the issue.
- I doubt Mike would block a technically flawless
  patch just because he doesn't like the author.
  If he tried to, he wouldn't succeed because I believe
  the other members of the developer community wouldn't
  let him get away with it. What you need to do, however,
  is to make the issue visible and comprehensible on the
  mailing list, and find others to back up your position.
  If you can't, then the problem isn't with Mike but with
  your patch.
- It is however obvious that Mike thinks that your proposal
  was inacceptable. Whether this is based on facts or
  if he just didn't get it I can not say.

> >I hereby offer to help review any alternative suggestions you may have
> >without bias or prejudice. I'll need other members to help.
> that's great and appreciated, so if you want a history of the previous
> discussions let me know.
> There you'll also see that any review from Mauro end up in an accepted
> improvement, this just proves that other ways are possible too.

Hans Verkuil also offered to help:

So there is still hope to get this issue fixed eventually.

OK, lastly I have to say that in the wake of all this
a nasty finger pointing game and spreading of FUD
in private mails enfolded behind the scenes, which
I find pretty distasteful. A couple of people told
me that they can't trust anyone anymore, and while
I can't rule out that one or the other actually enjoys
playing such FUD-spreading games, I think the root of
he problem is that some people take themselves too
seriously, and are too unforgiving about the quirks of
and mistakes made by others.


More information about the linux-dvb mailing list