Re: to be effcient or not to be efficient

Attila Kinali wrote:

>>I mostly agree with you arguments but would like to add yet another:
>>- Robustness on data corruption/missing data.
> 
> 
> This is something the demuxer and codecs have to handle, it
> has nothing to do with the API itself.

It does.
Let's consider you have a pull model. What happens if you pull data for 
timecode X and finally you get just an error ? What happens next ? You 
pull data for what timecode ? (I assume you use timecodes)