[FFmpeg-devel] [PATCH] [2/??] [3/3] Filter graphs - Parser for a graph description
Fri Mar 21 16:50:20 CET 2008
On 18 Mar 2008, at 23:12, Michael Niedermayer wrote:
>> So IMO, this case just show that "mathematical" description allows
>> to represent complex graph as well as the original description.
> no, it does not, it shows that you could represent the given graph.
> not a proof that all graphs can be represented. Though its quite easy
> to proof that if you want ...
Indeed all finite graphs can be represented, it's an algebra entirely
axiomatisable, no problems.
> More complex ones without feedback could look like
> i think first wins
It's mostly a matter of taste, to me they are equally complex because
in the first form I may struggle to follow the threading; but as I
wrote in a previous message you need to refine you language to specify
the order of the outputs out of a complex filter.
> With one feedback
Yes, feedback is tricky, no question. But with names you either have
to accept that each time you need a particular feedback you have to
hack the code of your filter (see my previous message), or you have to
provide a some combinator to do it.
> And with multiple feedbacks we have seen that the !*, description is
> significantly less readable
Agreed, multiple feedback should only be handled with some carefully
> So IMHO while the * operator might be worth considering, the way
> are handled is absolutely messy when there is more than 1 feedback.
We have a problem of terminology here: while I agree is difficult to
follow, and clumsy to write, and that there might be better ways, I
don't agree it's messy: it is very precise and well defined, it can be
explained in one line, no further confusion is possible.
> Also i think ,* is unintuitiv |, might be better but then maybe not
> i dunno
Not sure I understand, I expect "*" is the same as Vitor's ";" , just
a different notation.
More information about the ffmpeg-devel