On Friday 20 Mar 2009, Siddhesh Poyarekar wrote: > On Fri, Mar 20, 2009 at 11:40 AM, Dinesh A. Joshi > > <[email protected]> wrote: > >> When everybody elses hardware works and yours doesnt, It means > >> flaky hardware. > > Yes, you will probably have a clue about how this works when you > get into how exactly some hardware work (or don't). Recently I came > across an issue where a certain DVD drive lies about the type of > disk it has and we had to hack in a fix for it. That in my opinion > is flaky hardware, not a driver issue per say. > > Why do such hacks take time to come into Debian sid? I suspect its > because of their QA process,
IMO the real reason is that the identification (manifestation) of flaky hardware depends on several factors quite difficult to reproduce. Add things like "INTEL hardware" with jmicron ide + vodoo blob firmware and you have a PITA. Unless you check the chip revision numbers on the board you really dont know what you are buying. There are many other aspects too, related to the present methods of mass assembly of mobos and physically mounting them within boxes. The stress tolerances on a particular ball of a device depend on .5 mm positioning accuracy in a 3d plane and the torque on the screws. And dont go by a brand. They usually have crappy work arounds plastered inside the doze drivers. If the board actually proves to be popular the 2nd or 3rd revision might solve some issues and introduce some more. The answer to a hardware + software question is valid only for a very tiny set of well defined conditions. Recently Rony and I had a problem of bad sata power cable. Rarely intermittent. Would make one suspect the disk. So is Debian sarge more stable than Ubuntu hairy rhino? Cannot say without running both on the exact same hardware. The point i was trying to impress without getting DJ undies in a knot. -- Rgds JTD -- http://mm.glug-bom.org/mailman/listinfo/linuxers
