Re: [CinCVS] Re: Fwd: [estudiolivre] I believe in cinelerra --cin2 or cin3?

On Sun, January 20, 2008 17:44, Herman Robak wrote:
> On Sun, 20 Jan 2008 07:39:55 +0100, Christian Thaeter <ct@xxxxxxxxxx>
> wrote:
>
>> Fabianne Balvedi wrote:
>>> ---------- Forwarded message ----------
>>> From: Leo germani <leogermani@xxxxxxxxx>
>>> Date: Jan 18, 2008 7:53 PM
>>> Subject: [estudiolivre] I believe in cinelerra
>>> To: estudiolivre <estudiolivre@xxxxxxxxxxxxxxxx>, Felipe Fonseca
>>> <felipefonseca@xxxxxxxxx>
> ...
>>> First thing to do is gather this people to discuss about the future of
>>> cinelerra, identify the main flaws and its solution, make a plan to
>>> organize the place and set up for new features.
>>>
>> I dont want to go into the details, just an introduction:
>>  * We rewrite it from scratch, but reuse code and ideas where possible
>>  * We work bottom up, first a core render engine, the GUI at last
>>  * Being language agnostic, plugin-interfaces will be plain old C then
>>    it is possible to write things in different programming languages
>>  * Using a free distributed devlopment model, that is:
>>   * for the developers, there is no (mandatory) central point, anyone
>>     can work on it, everyone is equal
>>   * Discussion is done on irc or in private mails
>>   * final decisions are published inside the repository docs
>>   * lowering entry barriers as much as possible
>
>   This is already taking place.  However, it is not an effort to patch
> Cinelerra 2.1; more on that below...
>
>>> 2. Diagnostics
>>>
>>> Cinelerra code is not very well documented, so few people have the
>>> idea of how tuff is to deal with it. Second step is to see what must
>>> be done so we can invite more people to colaborate with the code.
>>> Documentation, refactoring, etc. It also has to work on the API so
>>> other people can write plugins and effects.
>>>
>> This is one, albeit minor of the reasons why we want a rewrite for
>> cinelerra3. There is too much cruft which makes refactoring (i tried
>> quite a lot and gave up) hard and unsatisfactionary. Next, since
>> upstream doesnt cooperate very well the only gain we get is a refactored
>> sourcebase which lost the mergability with upstream but still much old
>> undocumented design cruft.
>
>   Let's be clear about the options:  Cinelerra3 will take 2-3 years to
> produce something suitable for end users.  Even with a massive influx
> of skilled coders there will be no production quality Cinelerra3
> _this year_.
>
>   Patching Cinelerra 2.x to an acceptable level _may_ be doable this
> year.  I think the statement "I believe in Cinelerra" was a vote in
> favour of patching Cinelerra 2.x, otherwise it should have said
> "I believe in the future of Cinelerra".
>
>
>   What to do?
>
> 1) Compete fiercely to get as many coders as possible to join
>   either camp (Cin2 or Cin3)
>
> 2) Decide a parallell roadmap where Cin3 is poised to replace
>   Cin2 _when it's ready_, and Cin2 is maintained until then.
>
> 3) Let the developers send in patches when they feel like it,
>   and see where things are heading (roughly the current state)
>
> --
> Herman Robak
>
> _______________________________________________
> piksel mailing list
> piksel@xxxxxx
> https://www.bek.no/mailman/listinfo/piksel
> http://www.piksel.no
>



I keep asking the question: why Cinelerra ? Why not get involved in LiVES
development, it doesn't need rewriting from scratch, it just needs more
developers to get involved in using what's there now - mainly coding
plugins, and perhaps a couple of developers to make changes to the core
(if there is anything that needs adding - I guess the main thing LiVES is
lacking right now is faster/instant file import on the first load). A user
manual would be nice, and maybe an artist to design a new
theme/splashscreen for the 1.0 release. Cinelerra had a few years
headstart, but LiVES has been catching up fast.

Cheers,
Salsaman.
http://lives.sf.net


_______________________________________________
piksel mailing list
piksel@xxxxxx
https://www.bek.no/mailman/listinfo/piksel
http://www.piksel.no