On Mon, 13 Jan 2003, Mike A. Harris wrote:
> On Sun, 12 Jan 2003, Vladimir Dergachev wrote:
>
> >> I'm very happy to hear this! I hope we can get more of a developer
> >> community going, like I've been feeling in #xfree86 recently. However,
> >> I still think the specs are valuable (I've used a couple of them, and
> >> expect to again) even if the current set is limited. Are there any
> >
> >I quite agree. The fact is that without hardware specs there is
> >truly little reason why someone would be interested to working
> >with XFree86 code instead of writing a userland app. Granted,
> >XFree86 offers network transparency, but why bother working
> >within Xserver for a prototype app ? And people who do not get a
> >chance to work on a toy project within X tree are much less
> >likely to contribute.
>
> Where hardware is concerned, such as the video drivers, I would
> agree. More often than not, having the hardware documentation is
> required, or is at least quite helpful. However, the entire
> XFree86 sources contain libraries, extensions, applications,
> fonts, documentation, and many other things - all of which
> other potential developers could volunteer to work on too,
> without requiring hardware documentation.
>
They could volunteer to work on it, yes. However a good many of open
source efforts starts with need to scratch your own itch. All I am saying
is that hardware-independent itch can be scratched without messing with
Xserver internals.
>
> >> One thing that I think would help a lot here would be more feedback on
> >> what was lacking in our patches. I know I submitted some broken ones to
> >> you, but when you later fixed them I didn't get a note saying "Hey, we
> >> support FreeBSD back to release X.X, please respect older versions."
> >> Anything to tell me more of what is expected for the project will help
> >> me submit better patches in the future.
> >
> >There is also an issue of Bazaar versus Cathedral type of development.
> >There is a tradeoff between quality of patches and number of people
> >actively working with the code.
>
> I'm not sure what you're suggesting here. Could you clarify a
> bit?
>
> The Linux kernel for example has a very large source code base,
> and has countless developers whom have worked on it under the
> Bazaar model, and the code is quite high quality. People write
> good patches, and people write bad patches regardless of what
> model of development is used. However, I think that due to the
> bazaar style of development the kernel is done under, the bad
> patches get weeded out much more easily, whereas with the
> cathedral style of development, they're more likely to simply get
> ignored or cast aside.
>
> Is that similar to what you mean?
This seems to be another angle ;) My point was that the easier it is for
patches to go in the more attractive the project looks to new developers.
best
Vladimir Dergachev
>
> TIA
>
>
> --
> Mike A. Harris
>
>
>
>
> _______________________________________________
> Devel mailing list
> [EMAIL PROTECTED]
> http://XFree86.Org/mailman/listinfo/devel
>
_______________________________________________
Devel mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/devel
