Date: Fri, 12 Jun 1998 08:14:00 -0600 From: Wes Peters <wes@softweyr.com> To: Jack Velte <jackv@earthling.net>, freebsd-chat@FreeBSD.ORG Subject: Re: gartner group article Message-ID: <358137A8.9CE61701@softweyr.com> References: <01bd9517$c4f949a0$1001aace@eliot.pacbell.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Jack Velte wrote:
> a serious FreeBSD would hurt sco, hwp, sunw, and msft.
It's more likely to help HP and Sun. Once you install some FreeBSD servers,
and start to learn what a GOOD system unix really is, when you need to move
beyond the performance offered by FreeBSD, what're you going to go to?
Neither SCO nor M$ offer *better* performance than FreeBSD, they both offer
*less* performance on the same hardware. So, you scale up to Sun and HP,
onto the machines where they have the most margin. Their low-end servers
and workstations don't offer more performance than FreeBSD, they offer
less performance for greater prices.
> >While the
> >development of SMP support in -current has been herculean
> >that even our intrepid -core members will have the time or machine
> >resources to tune FreeBSD-SMP for machines with 64, 128, or 256
> >processors.
>
> even if they were paid $400/hour?
Not unless you can:
a) get somebody to make such a machine,
b) document it sufficiently that somebody else (i.e. FreeBSD Inc.) can
bring up an SMP or NUMA architecture OS on it,
c) buy one of them,
d) let the MP guys crash the hell out of it for the next 6 - 12 months.
For item a, Sequent probably has a 32-processor machine that might fit
the bill. Having been familiar with their OEM support in the past,
they might even fit "b" iff they're willing to give you access to their
source code, bus arbitration code, etc., and their internal support
team. Not likely. Items c and d aren't likely at all, unless you
have a million or so lying about not doing anything, in which case you're
a fool, because you could've spent it on a boat and just forget all this
computer silliness. ;^)
> >Their conclusions about NT in this arena are colored a bit by
> >Microsoft marketing hype. The 'trade press'
>
> at MIT's business school, they teach NT is 100% better than anything else.
> but it's such a fraud.
The first myth of management is that it exists.
MIT has been in the business of pumping out dickheads who don't know
business processes from their asses, and get stuffed into high-
paying no-productivity jobs by the good-old-boy network for 300
years. This is the pack of morons that came up with "matrix management,"
after all. You can't expect them to change just because of some little
burp in the economy like computers.
> > seems far too willing
> >to believe Microsoft has tamed the 'thousand monkeys' approach to
> >developing large software projects. In fact, I'll go so far as
> >to coin "Wes Peters corollary to Brooks Law:"
> >
> >"Adding thousands of programmers to a slow program makes it slower."
>
> how many people actually work on the FreeBSD operating system?
Rarely more than 3 to 5 on any particular feature, and often each area
of expertise is covered by one particularly competent programmer,
sometimes with the help of a few other accolytes. FreeBSD falls some-
where between the Cathedral and the Bazaar (or Bizarre, in the case
of Linux) in team philosophy. In Yourdon-speak, FreeBSD uses multiple
"operating room teams" under the loose coordination of "the core team".
(From the outside, it seems like Jordan-the-octopus does a lot of the
coordinating, but that's an outsiders point of view. The inside view,
as always, is probably greatly different and much more bloody.)
Somebody oughtta "right" an article about this. ;^)
--
"Where am I, and what am I doing in this handbasket?"
Wes Peters Softweyr LLC
http://www.softweyr.com/~softweyr wes@softweyr.com
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?358137A8.9CE61701>