begin jfweber@bellsouth.net's quote:
| *** Reply to message from dep on Wed, 30 Apr
| 2003 19:48:08 -0400***
|
| > if, however, the purpose of the exercise is getting
| > suse running on customers' machines, then acpi defaulting to off,
| > with documentation describing how to turn it on, would seem to be
| > the right thing to do, don't you agree?
|
| It might be, if SuSE also ran a hardware division , but there just
| aren't enough guys and gals there to monitor everything. Perhaps a
| lot of newer boards require it ? THAT would really put a crimp into
| the adoption , if someone , who may be trying to get the company to
| look at Linux as a solution discovers too late ( and w/ the big
| boss watching) that the test box wont boot at all until someone
| realizes that they need the aspci on in order to work.
if someone demonstrating for the boss hasn't booted the frigging thing
to see if it will work ahead of time, that person is doomed to get
fired before long anyway. and i am not proposing that anything be
utterly disabled -- i'm proposing that the default and the
documentation be changed. no one is supposing that suse can do all
possible hardware tests; indeed, the fact that the can't is precisely
my point -- they cannot, therefore, make presumptions about hardware.
| IF it's true that newer hardware ( say the last 5 years or so )
| needs to have it on, it's not a bad guess to have it default to
| on.. w/ relatively easy fix ( emphasis on relatively) to get the
| older ones up and running. OF course Linux sysadmin has probably
| read a lot of the pre install info so isn't flustered. OR he knows
| his boxen need the thing off or on anyway. Can't be said for *most*
| of the w32 sysadmins I've met tho . <g>
it isn't true of newer hardware -- the machine on which i had the
difficulty includes a motherboard that was new three weeks ago and
was released in january, and it will not work with suse's acpi. nor
will one a year old, nor three that are in the neighborhood of two
years old.
in that this has been a problem for a substantial number of users,
making it user-configurable early on, as well as documenting its
failure modes (these include loss of network card and other pci
devices) such that installers would be drawn to the non-obvious acpi
problem, seems at a minimum prudent. i do not see why there is so
much resistance to this.
--
dep
http://www.linuxandmain.com -- outside the box, barely within
the envelope, and no animated paperclip anywhere.