Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 20 Aug 1998 10:21:59 -0700 (PDT)
From:      Ken McGlothlen <mcglk@serv.net>
To:        malte.lance@gmx.net, questions@FreeBSD.ORG
Subject:   Re: Floating point exceptions on i386 and FreeBSD-2.2.X
Message-ID:  <199808201721.KAA17051@ralf.serv.net>
References:  <199808180818.AA14374@hilbert.wias-berlin.de> <13786.45438.762966.929694@neuron.webmore.de> <199808191806.LAA12320@ralf.serv.net> <13787.55346.810743.132179@neuron.webmore.de>

next in thread | previous in thread | raw e-mail | index | archive | help
malte.lance@gmx.net (Malte Lance) writes:

| Ken McGlothlen writes:
|  > That is about as inappropriate a response as I've seen on a public support
|  > list, ever.  Whining about perceived lack of features or bugs is actually an
|  > appropriate, if occasionally annoying, activity on the part of people who
|  > encounter things they don't expect on a system.  Saying that they look like a
|  > little pussy is completely out-to-lunch, and does a disservice to the list, and
|  > to the perception of FreeBSD.

| First of all Mr. Griepentrog said in his post, he was using FreeBSD for 3
| years and was very satisfied with it.  Then in the same paragraph he tells us
| about his recent disappointment with FreeBSD. [...]

First of all, I would like to point out that I understand some of the
frustrations of tech support, particularly when it involves a product you have
some personal investment in and considerably enthusiasm for.  Believe me when I
say this:  I live in Seattle, the heart of Microsoft Country, and for around
95% of the businesses out here, NT Is The Best Thar Is.  They're phasing out
their Suns, their Macs, and pretty much everything else to go NT---and they're
either baffled that their support costs are spiraling, or completely ignorant
(at the moment) that they are.  Just *try* pushing FreeBSD in this environment.

[Oddly enough, one of the best tools I have is just setting up a FreeBSD box
and an NT box, running on identical (and I mean *identical*) hardware, and
letting them run their respective screensavers.  My favorite ploy is to set up
the NT "Stars" screensaver, with 200 stars (the maximum) on a relatively fast
speed.  It looks terribly jerky.  Then I set up

	xtacy -root -few -stars -number 2048 -delay 0

on the Unix box in xscreensaver.  I had to increase the number of stars to 2048
just to SLOW DOWN the box enough.  I also keep -number 200 around just for
reference.  Dang if that FreeBSD box don't look tons faster.]

But I still think that the appropriate method for dealing with misperceptions
(and that's all this was, was a misperception) is to correct the perception,
and do it in an engaging and preferably hypertolerant manner.  Believe me:
they'll feel more stupid than if you'd attacked them, but the won't feel like
you hate them.  One example would have been in this case:

	Wellllll, I can see how you'd get that impression, because what Linux
	does (in its infinite wisdom) is mask the floating-point exceptions by
	default altogether.  So you still get inaccurate results (you should
	check that out sometime), but you just don't get *told* about it.
	FreeBSD is just being honest with you.

	Fortunately, FreeBSD has a solution, already in use by mathematicians
	worldwide.  Check out "apropos float" sometime, paying special
	attention to fpgetmask(3), fpsetprec(3) and fpgetprec(3).

A little smug, perhaps, but gentle.  In one fell swoop, you deprecate the
"Linux does better" feeling, and let them see that the answer was right in
front of them the whole time.  The main downside is that it takes a lot more
time to answer questions this way.  :l  (Not a trivial concern to people like
you, and especially Doug.)

In any case, I do empathize, and I'm sorry he got to you like that.  I trust
you'll be able to take a deep breath next time.  :)

							---Ken

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199808201721.KAA17051>