Hardware regressions was (Re: Do you really want developers to be on this list was (Re: Very bad status of hardware (especially wifi) support in ubuntu, due to the too many accumulated regressions))
ubuntu at kitterman.com
Thu Nov 13 11:35:26 UTC 2008
On Thu, 13 Nov 2008 09:00:36 +0000 Vincenzo Ciancia <ciancia at di.unipi.it>
>> If a distributor adds more goodies to the kernel, then be happy, but
>> that doesn't mean, that it really works...even when the distributor puts
>> the hardware on the list of supported hardware.
>I hope this is not really the idea of the ubuntu developers on this
>topic, because if so, then I can really, really forget all my bugs, and
>go home happy. If the idea is that a trial-and-error process should be
>the normal way of using ubuntu (it is the way I use it every time I
>install it to other people), then just tell me. I think it's
>unbelievable how far things went in this direction. If this is
>considered normal and unharmful, there's clearly something that I didn't
Part of what goes on is that the details of a product change over time,
where a specific part was made, or any number of things. So when one
person says (to pick one example, this is true for all vendors) IW 3945 is
broken and another says it's not, they probably don't have identical cards.
We also have more than one kernel. Maybe it works with i386, but is broken
Use cases differ too. I have a laptop with IW 4965 and it works great for
me. A lot of people reported problems on Intrepid with this card. As it
happens, I am mostly (maybe always) on 802.11G networks. People with the
problems have 802.11N (mostly anyway - see the other factors).
Part of the trouble with a hardware database is what to put in it to make
it reliable, yet searchable. So this is not an easy problem. Back in Edgy
I remember spending a lot of time digging through wiki pages trying to
figure out wifi. Clearly we need to do better with this, but I'm not sure
I think this may be a topic to take up with the QA team.
More information about the Ubuntu-devel-discuss