[XviD-devel] Changes in Source <-> Xvid 1.0

[Christoph Lampert]

Hi guys, Michael, Radek, Edouard,

all the three of you are right. I think what Michael wanted to express is
_not_ that you, Radek and Edouard, aren't doing a great job at making XviD
better and more stable. The mail could have been much shorter if he had
wanted to say that. I mean that, especially since you are by far the most
active people at XviD at the moment, while Michael, Pete and I all seems
to be rather busy with other stuff at the moment. 

I see the thing as follows: The question is "What is XviD 1.0 supposed to
be?" Michael wants to release, soon, the codec with the features it has. 
It's supposed to be rock-solid and bugfree, where bugfree means "Free of
noticable bugs for the users" where emphasis lies on stuff like image 
problems, non-compatible bitstreams, those things. 

Not being threadsafe is a bug, too, no question about that, but XviD has
never been completely threadsafe, and most users wouldn't call it a bug.
Rather a "missing feature", or they wouldn't notice at all. And this is
just meant as an example. "Beautifying" code is something else, that is
surely wishable, but it undeniably has the danger of introducing small
and annoying errors. Most likely, any errors are easily spotted and
removed, as in ME, but it is also a question of minimizing the risk. 
Hoever, as you mention, it also has the potential that errors are detected
and removed during that process, you are right about that, Radek. 

It boils down to a question like this: Which bugs do we really want to get
rid off before releasing? We'll never manage to get rid of really all
bugs. And I agree with Michael that XviD should be absolutely free of MPEG
errors, and visual problems. I myself would be furious if I got a
longexpected version 1.0, encoded several hours of video and later found
out it was all in vain. Other, less severe errors, non-working features,
inconsistent dialogues, non-optimized features or all such stuff is bad,
too, but usually there are workarounds until a new version appears.

We all agree that XVID 1.0 should be released, and released soon. I think
we all waited too long for this, and since there wasn't anything
"official" to do, started looking at things to make better. Which we did.
But I don't think, the recent changes are that important such that they
justify to continue working in this way and not release. 

* Let's have everybody finished with what he is working at. I think we all
trust each other enough such that when somebody says the code is tested
and most likely bugfree, we believe that. This should be possible within a
few days. 
* Let's create a public beta. It doesn't have to be the finished version,
it can e.g. have MMX-code missing, or whatever. Just let there be a
version which is really superior to everything we released to far. 
If bugs turn up in old or new code, especially of the image/MPEG kind, 
then even better, so we can squash them together. 
* Let's use feedback from beta to make a release 1.0. Soon. This year!
* Then, let's all get drunk at my place. I'll send you the invitation by
private mail (Airport "Cologne/Bonn" is very close, several cheap airlines
go there, starting at 19 Euros per flight). 

Okay?

gruel