Date: Sat, 27 Sep 1997 23:02:57 -0600 (MDT) From: Wes Peters <softweyr@xmission.com> To: Greg Lehey <grog@lemis.com> Cc: chat@freebsd.org Subject: DNS vs. GUI? (was: Microsoft brainrot...) Message-ID: <199709280502.XAA21135@obie.softweyr.ml.org> In-Reply-To: <19970928101941.03210@lemis.com> References: <19970927143934.ZN26834@uriah.heep.sax.de> <199709272127.OAA11524@usr08.primenet.com> <19970928101941.03210@lemis.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Terry Lambert opined:
% UNIX users are, as a class, intellectual elitists who don't
% undertand that the average I.Q. is 100 because that is how a 100 I.Q.
% is defined. And as a class, they are unprepared to make the necessary
% allowances. There's a good reason a moron can run Microsoft OS's: so
% that that morons won't be too intimidated to buy them.
And this coming from Terry. Amusing. I think you're dead wrong, Terry.
I see UNIX users, as a class, as intellectual elitists who completely
understand that the average I.Q. is 100 because that is defined as
average, and refuse to let this fact interrupt their pursuit of
"better." Am I supposed to tie 76 IQ points behind my back every time I
sit down at my computer? ;^) Anyone who thinks a computing platform
designed for the average U.S. office worker is also suitable for
advanced software development, serving web pages, and running games is
an absolute idiot. That's like saying that a Pinto station wagon is the
perfect wheeled vehicle for picking up groceries, delivering sand and
gravel to the cement plant, exploring the surface of another planet, and
racing in a grand prix.
Greg Lehey writes:
> Well, again I'd say wrong on both counts. Morons can't run
> Microsoft's OSs. Even people of normal intelligence (whom I think I
> can understand quite well) feel intimidated by them. That's not to
> say, of course, that they don't feel intimidated by UNIX as well.
Er, yes. Exactly. There seems to be this common belief that since
Virus95 and NT are both "just Windows" that anybody can administer
and/or develop software for them. Bzzzt! Amazing how many horror
stories there are in the industry based on these two commonly held
beliefs. It's even more amazing how difficult it is to change the minds
of the dipshits who made the mistake in the first place! Let me
iterate:
1) Administering a farm of NT Workstations is more graphical than a
similar number of UNIX workstations, but no less work. In fact, it's
quite a bit more work, because Microsoft has done nothing to make it
possible to administer them as a whole. I understand this is changing
significantly in NT 5.0, and it's about time. UNIX has a few tricks up
it's sleeve in this area, but not nearly enough. NIS was conceptually a
step in the right direction, but is a poor architecture to rely on.
2) Assiging "Windows" programmers to write server applications is a sure
recipe for failure. If you are developing network server code,
regardless of the platform, you need someone who understands the realm
of network servers. This does not generally include someone who learned
to program using VC++ (or even Turbo/Borland) and thinks of a network
connection as a slow disk drive.
[much argument about setting up DNS via a gui ensues...]
% It's a hell of a lot easier than running an editor and
% shelling out to a man page every five minutes, or worse, shelling out
% money for an O'Reilly book, which is made necessary by all the bogus
% complexity associated with the task, and trying to balance the thing
% on your knees while typing in secret code words.
HOW MANY TIMES DO I HAVE TO REPEAT THIS HERE?
You *all* *STILL* seem to have missed the point. The problem with the
Microsoft GUI is not that they're using a GUI for a complex task, but
that they're using the GUI *poorly*. All they've done is stick a couple
of really stupid dialog boxes on top of the configuration file, which
doesn't really add any value.
In case you wonder why I'm shouting, WE JUST HAD THIS DISCUSSION (vis a
vis users and groups) A COUPLE OF WEEKS AGO! Now listen up, children:
The "standard" DNS server setups are organized by record type, because
that is how DNS was developed. The record types evolved over time, new
ones were added, but everyone seems to have missed a critical point
here: what DNS *really does* is describe some *OBJECTS*. In the case of
DNS, the objects described are DOMAINS and HOSTS.
Now, why don't we trash all of the record types and listboxes and such
and design a DNS configuration tool that displays these in a meaningful
manner? Including a reasonable inheritance mechanism, so that when you
place host "foo" within domain "bar.com", if no "Mail Exchanger"
attribute is specified for "foo", it will inherit the one from
"bar.com". Need I mention that once you're done, you can probably sell
it for quite a bit of money?
(Hmmm. This might be the application I've been looking for to *really*
learn Java. Where's that Visual Cafe, dag nabbit?)
> To be fair, I think that O'Reilly's DNS book is too confusing. TCP/IP
> Network administration will give you more info that any Microsoft toy
> config tool can.
Yeah, but if you really wanna see a neat DNS server, check out the
"Instant IP" feature in the Internet Station. Talk about zero
administration: the DNS server lets the workstation tell it what the
"hostname" is:
http://www.dayna.com/dayna/products/istation/instantip.html
% For something designed by a bunch of bonifide computer scientists,
% you'd think they would be able to grasp the concept of putting
% configuration databases in third normal form. 8-|.
> Who are you talking about here?
Database theory. Something Terry knows just enough about to shoot at
people who know *nothing* about it. If Terry knew a little more about
database theory, he'd realize that almost *no* working databases get
past the second normal form because nobody but database experts can
spend enough time to work the entire database into the third normal
form. ;^)
--
"Where am I, and what am I doing in this handbasket?"
Wes Peters Softweyr LLC
http://www.xmission.com/~softweyr softweyr@xmission.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709280502.XAA21135>