CD integrity checking (Re: Colony 3, Test report)

[Matt Zimmerman]

On Sat, Aug 20, 2005 at 01:57:15PM -0400, Alvin Thompson wrote:

> all right, you obviously aren't willing to be convinced so i'll drop it.
> i will point out that you could probably detect 90% of bad CDs in 1/10 of
> the time of a systematic scan by checking only the likely 'hotspots'
> (padding, etc). and the perceived user experience of a bad install is
> extremely negative, even though it's technically not the distro's fault.

I have tried to explain to you, with specific technical examples, why I
don't think your proposal is the best way to solve the problem.  If I didn't
think the subject was worth discussing, I wouldn't waste my time writing
email about it.

Things that I am not willing to do:

- Add a new, mandatory 8-20+ minute step to the start of the installation

- Engineer an overcomplicated parallel processing solution (and associated
  UI) to try to do an integrity check at the same time that the installer is
  running (which I don't think would work very well anyway)

If you have an idea for a check which takes 10% of the time while achieving
90% of the results, I would be very interested to hear concrete, verifiable
details about it.

-- 
 - mdz