hardware agnostic?

[Derek Broughton]

Muzer wrote:

> Derek Broughton wrote:
>>>>>         
>>>> I wouldn't be _shocked_, but I mean that Scott's plans, heavily
>>>> dependent on udev, involve simply _not_ loading any software that is
>>>> not required to get
>>>> X up and running.  Everything's geared to making X start as fast as
>>>> possible.  So while you might be able to achieve the same thing by
>>>> compiling a kernel for specific hardware, you can make it _look_ as if
>>>> everything is loaded just as fast by waiting for the really intensive
>>>> applications to
>>>> settle down before you load other things.  I guess I really wasn't "not
>>>> convinced" that you could speed up boot by compiling the kernel, but
>>>> that I wasn't convinced it was either the only, or necessarily even the
>>>> best, way.
>>>>   
>>> He says "end of bootup" is when both CPU and I/O is idle.
>>
>> Who?  Scott?  I've seen him say "when you have a working desktop" - and
>> whether or not that's strictly the "end of boot", it'll probably do for
>> most people.
>>
>>   
> https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2009-
June/008505.html
> 
> "And just to affirm something we've already stated; this benchmark time
> is to a fully logged in desktop (auto-login) with an idle CPU and Disk.
> Deferring services is not an option unless done properly (ie. switching
> services from startup to on-demand activation)."

Ah.  I see your point, but naturally I interpret it differently.  He's 
talking about a boot into an absolutely clean desktop - no startup scripts, 
no saved sessions.  And, surely, it it reaches idle, you've finished 
booting.  But that's not really "normal usage", and he still makes the point 
that you can _properly_ make your services start on-demand, and says nothing 
_there_ about deferring drivers, but the very point that it all depends on 
udev, and udev prioritizing the modules for X, implies that deferring module 
loading _is_ important.
-- 
derek