[NTLUG:Discuss] Fry's and Linux
Kevin Hulse
hulse_kevin at yahoo.com
Fri Jun 18 17:10:37 CDT 2004
--- "Burton M. Strauss III"
<Burton_Strauss at comcast.net> wrote:
> Oh I can't resist...
>
> > -----Original Message-----
> > From: discuss-bounces at ntlug.org
> [mailto:discuss-bounces at ntlug.org]On
[deletia]
> > That gives you two massive advantages over
> whatever
> > process produced the manufacturers information: a
> > wider test sample and use under conditions
> comparable
> > to what an end user would tolerate.
> >
> > If someone was willing to work through the
> problems to
> > get something to work, chances are that it really
> does
> > work.
>
> Yes, and no. The problem is the vastly different
> levels of skills
> available. And the difference in systems - two
> years down the road, my
> Dad's Dell is almost untouched from what left the
> factory. But only Dr.
> Frankenstein would recognize my Brother-in-Law's one
> year old Dell...
>
> Here's what happens.
>
> User A, a newbie puts card B into system C and finds
> it doesn't work. Posts
> a question, gets no help, gives up. Reports it does
> not work. Takes it
> back to Fry's who marks it down $5 and puts it back
> on the shelf.
This would be one level of disfunction. This would
merely mean that kudzu (or the pnp daemon of the
moment) was unable to handle the device at the
highest level of automation/ease.
This is easy enough to express in a compatibility
matrix.
>
> Vastly experienced user Z, puts the same card B into
> a system C' (very
> similar to C), sets the parameters in
> /etc/modules.conf and finds it works
> great. Reports that it works, but doesn't mention
> the settings - he's an
> experienced user and this is normal for any card.
This is normal for any THING. This is a detail that
something like kudzu would hide from you. This is also
something that is dead simple if you have even the
slightest admin level of knowledge about Linux.
This is also easy enough to express in a simple
compatibility matrix.
>
> Now you go out looking for reports on card B+ in
> your FrobozMagicCarpet
> 200A74. There's a report out there for a Froboz200A
> says it doesn't work
> and for FrobozMagicCarpet 200A70 that says it does.
>
> You try it and it fails. Not because the card won't
> work, but because all
> of the information isn't available to you - that you
> need to tweak
> /etc/modules.conf...
Yup, kind of like certain USB & PCMCIA devices that
really get cranky if you plug them in before
installing the corresponding Win32 driver. A little
documentation goes a long way in such cases. It's in
the interest of people who want to keep your money to
provide such information.
Ditto for really weak wifi antennas, cheap dvd players
with poor compatibility or any number of other subtle
consumer issues.
>
> Is the information available to you 'wrong'? No.
> Incomplete? Yes.
> Out-dated? Maybe - what was true in 2.4.9 isn't in
> 2.4.18 for many
> devices...
>
>
> The differences can be incredibly subtle. Linksys
> makes (made) an 8 port
> 10/100 hub. But Linksys EFAH08W v1 isn't EXACTLY
> the same as a EFAH08W v3.
This is THE prime example of why merchants need to
have a more perfect understanding of what they sell.
All of the network vendors like to make dramatic
chipset changes within the same model number. Even
under Win32 this is bound to have meaningful
functional implications.
[deletia]
Build or acquire the kernel module before you even
think of buying the associated hardware. Some vendors
may go so far as to claim Linux compatibility on their
packaging when infact they only provide a binary
compatible with a particular distribution and kernel
version.
More information about the Discuss
mailing list