[Ffmpeg-cvslog] r5941 - trunk/doc/soc.txt

Robert Swain robert.swain
Mon Aug 7 22:09:07 CEST 2006


Michael Niedermayer wrote:
> if you have more people who want to work than people google is willing to pay,
> then you must select a subset of them and asking them question about C,
> optimizations, ... is one way to select such a subset of students
> if the number of students is <= the number google is willing to pay then
> of course should we accept everyone ...


> and the idea is not that this is a silly test for which you should sit down
> a week before and memorize a few things but rather a way to predict how
> good the final code will be, of course sitting down a week and learning
> stuff will have some effect on the quality of the final code but memorizing 
> stuff is easier then actually being able to apply that knowledge
> maybe we should give students a small exercise before the next SOC instead of
> some questions about C, maybe something like a messy piece of code which they
> then should cleanup and a buggy piece of code where they should find the bugs

I agree that a potential student being aware of and able to exhibit such skills 
would be a good idea. As I said they seem to be fairly simple things that you 
probably should expect but I don't think they're great rejection criteria 
because they don't take much time to learn even before embarking on the task.

> the big problem with any test though will be that some people will be nervous
> and will perform worse then they would if they are not nervous ...


> an alternative would be maybe that we could ask every student to show us some
> program (s)he wrote ...

When applying I provided all my source code to date, including a project written 
almost entirely by myself. I think this is a good suggestion but you may have 
concerns about cheating and the ownership of the work. It's a shame people can't 
be trusted to be honest.

I'm sure Mike would be able to comment better on the way the applications were 
reviewed but I assume the application gives a good idea of the overall 
suitability of a student for the task(s) for which they have applied. I don't 
doubt that had you made the selections (did you?) you would have been perfectly 
reasonable and that I probably took the notes about C knowledge a bit out of 
context unnecessarily.

For what it is worth, I would be happy to share my experiences with you to help 
mentors and students next year and I'm certainly going to hang around (though I 
was intending to anyway) to work on more things for FFmpeg in the future.


More information about the ffmpeg-cvslog mailing list