[MPlayer-dev-eng] [RFC] EOSD improvements

[Reimar Döffinger]

On Mon, Aug 02, 2010 at 12:14:47PM +0200, Grigori Goronzy wrote:
> On 08/01/2010 11:25 PM, Reimar Döffinger wrote:
> > I thought I said clearly the low-level support is _not_ my concern,
> >
> 
> IMHO it wasn't that clear.
> 
> > my concern is that we might end up with working but _useless_ low-level
> > support because the high-level stuff (time stamp handling etc.) might
> > mean it does not work for any of the cases we want to have it working.
> > Or to put it differently: I suggest to implement a as simple as possible
> > proof-of-concept that does something useful before starting with API-level
> > work.
> 
> This all boils down to design philosophy - top down vs bottom up, which
> are a matter of taste to some degree.
> Also, low-level support is not _useless_ just because it isn't used by
> some higher-level subsystem from the start. If the interface is good and
> functionality is interesting (and it definitely is), someone is going to
> use it.

That's the big if. IMO very, very few people are capable of designing a
good interface without having a heavy user right at hand.
I considered it a good idea to assume that you aren't among those few
people and thus suggest an approach that I think is likely to work better for
anyone with average skills.
And the low-level support _is_ useless if there never will be any high-level
user because the interface in the end happens to be just broken - doing
both at the same time is quite a reliable way to avoid it being
_completely_ broken at least.
Another problem is that without some actual "user" for the code, it
is hard for me to suggest simplifications to any patch or accept
some complexity as unavoidable.
I am telling you which approach I do believe will get quicker results,
also when it comes to getting the code reviewed.
I can't and don't intend to do more than give advice though, however I
do think I have made and fixed enough mistakes myself that it is a good
idea to at least listen and try to really understand what I say though,
even if you don't manage to agree.