[matroska-general] Re: Message to the C porting team

Steve Lhomme steve.lhomme at free.fr
Mon Feb 17 16:35:37 CET 2003


En réponse à Animesh Srivastava <animesh.srivastava at blr.techspan.com>:

Hi Animesh,

> I do realize the importance of the library, and I have been following
> closely the discussions on the mailing lists, - the website too
> mentioned
> the "impending" need of C library.. at the same time was trying to
> recode
> everything clean in C.. The way I was approachin it was fundamental -
> went
> thru the ebml specs of ours, and the wap ppl.. got the differences and
> fundamentals clear.. code EBML first.. and then see if porting other
> things in lib is also required by them.. 
> 
> Anyways,.. now that this impending is becoming really **immediate**,.

Well, as we also tend to say a lot, this project is based on people's good will.
We have to make things for fun, to learn, and to make this thing grow. I often
said that one of our advantage is that we are not a company, we have no
deadline. We can take the time to make things as good as possible.

In the other hand, there is a growing pressure on our project to have something
working for the public soon. Soon means that it's before the MPCF format and
Xiph Theora are going public. This is just a strategic deadline to have this
little user preference in the end (we all know technical superiority is not
enough). That's why we are all under pressure at the moment (and I really don't
like that).

I think once we have myFUN's Direct Show filter working with a basic file, we
can release all this alpha code to selected beta testers (as found on the
corecodec forum). Once we reach this stage, we will probably be able to work at
a normal development speed (people will have something to play with, comments,
rants, etc). And we'll have time to prepare the first public launch.

The MPlayer support could be part of this public launch (beta), but I don't
think it's a priority for this coming (soon) release. That means you can take
your time (not too much ;) to make something working well and that will please
them (and other coders).

According to Mosu (Moritz Bunkus) they don't like C++ for portability and
compilers portability issues. I can understand that if they share their
libraries with other apps (is that the case ?). And in the end they will want
something in C. Now a C wrapper of the C++ library will probably have all the
disadvantages of C++ and the ones of C. So I personally think that there should
be an alternative library (libmatroska_c) written in C only. This library could
be later used in low profile hardware players too.

So if you think you can handle the creation of this library in C, I think you
should invest your time in that way. You will sooner or later get more help from
other coders (including some of us) because more people know C well in the open
source community. You may even get support from the MPlayer people if matroska
becomes popular faster than expected :)

PS: This is my own point of view, others may have a different one.
http://matroska.org




More information about the Matroska-general mailing list