Re: Coaster-lib

[Manuel Clos <llanero eresmas net>]

Mike Messmore wrote:
> Wouldn't it be better to perform some sort of test to decide the fastest
> reliable speed a drive could write at froma  given source?  This would
> take into accound other factors and help prevent buffer underruns.  I
> think I've seen burning software for windows that does this, by testing
> how quickly you could burn from every hard drive and CD/DVD drive on the
> system.  This could be run the first time you want to burn something and
> then optionally run at a later time for people who change up their
> hardware.  The data could be stored with gconf so anything else that
> wants to access it could.  But then again I'm not familiar with the
> technical details of these tests so I may just be unknowingly talking
> out of my arse.

Hi Mike,

I really don't like these test. First of all, they annoy you the first 
time you want to burn a CD. These tests are not reliable. For example, 
the first part of the hard disk is faster than the last one, so which 
one will you test? perhaps the user won't never use the last part of the 
hard disk. Also, when burning just downloaded movies, programs like 
mldonkey, create these 700 Mb files with more that a 50% of 
fragmentation, which really hurts when reading the data. And you don't 
want to burn at 4x with your 16x drive.

The best solution is burnproof. When it is not available, the program 
can warn the user about the burning speed, just that.

--
Manuel Clos
llanero eresmas net

TCPA y Palladium: http://bulmalug.net/body.phtml?nIdNoticia=1398
TCPA and Palladium: http://www.cl.cam.ac.uk/~rja14/tcpa-faq.html