MAC vs. UNIX part 1
In the following I consider Brian Fristensky's commentary on
the mac vs. unix OS. I found it an good summary of the
problems with the mac, and advantages of unix in similar situations,
(especially for programmers). Since much biosci software is not
developed professionally, this is an important consideration.
I do have, a problem with his discussion of what should be expected
of a computer user. This is a somewhat philosophical point,
and I will come to it in another posting.
Good points(what's wrong with the mac):
b)Sequence Format Interconversion:
>What if you have a sequence in one format, and the next program you want
>to use requires a different format? (Say, a 1 or 2 at the end of the
>sequence to indicate topololgy.) Under Unix, you enter the vi (or other)
>editor, jump to the end, add a character, and quit. In Macintosh, you have
>to traverse the directory tree to your copy of MacWrite (or whatever),
>click on the cute icon, find your file by clicking through the filelist, import
>your ASCII file, make your change, export your file, and click back through
>the directory tree to where your datafile is. The point is YOU CAN'T MAKE
>QUICK CHANGES IN DATAFILES.
This is the thing I dislike most about the mac. Apple provided
a freely distributed editor, Edit for working with text files
(so one doesn't have to use import/export functions), but jumping
around directories is quite irritating. Text editing on a row/column basis
has deliberately been discouraged. This is probably a good thing for
a desktop publisher who found this model for text manipulation
constraining, but is a problem for us, whose programs expect formatted
strings of characters.
>PROGRAMMING. Because of the need for rapidly making changes in programs,
>specifying compiler options, linking, Make-ing etc., you can't just buy a
>compiler for a Macintosh. Instead, you have to buy (and learn how to use)
>something like the MPW Workbench, which creates what is essentially an
>internal operating system in which you do all of the programming steps. In
>cases in which your program is too big (ie. most of the time) you can't
True, MPW is a (partial) unix shell. However, most C programming on the mac
is still done using LightSpeed C, which is a 'native' mac compiler following
the interface, and its very fast. Even MPW has a bunch of canned tools
that allow you to go through the compile/link cycle without ever typing
a command. Memory simply isn't a problem unless you're programming on a
Plus. I only know of one programmer who has less than 2.5 megs, and most have
5 or 8.
The big mistake here was that Apple didn't offer any development tools on
the mac for almost 3 years. Programs were written on Lisas. If they had a
vision for development "the macintosh way", they should have shown it to
>Again, I pose the question: Why did Macintosh have to eliminate all of the
>features that make an operating system truely useful? They had to have had
>these things in-house, or they couldn't have written the thing in the first
>place. They DELIBERATELY make it as hard as possible to use anything other
>than what has been specifically provided.
The reasoning I have gotten from Apple on this is that if less
intuitive, more powerful ways of doing things were available, programmers
would write programs incorporating them! If it was easy to create a
command-line interface on the mac, there would be a lot of mac programs
that used it. Thus, new users would be confronted with a mixture of
program interfaces, some easy, some much harder. If you are selling a
lot of programs, high volume, to business, you can afford to make everyone
create easy to use/difficult to program interfaces. However, this is
a liability for those developing software for a small,nonpaying market.
Its even more extreme this. Apple rides herd on their "certified
developers", publishes a big volume of "user interface guidline",
and if you don't follow them you're "not of the body", etc., etc.
It turns out that fast editing of text files, file browsing, quick
launches of frequently used programs, etc., are all possible
through third-party software. Most "mac power users" (great concept!)
have a bunch of INITS/CDEVS in their system that give them these
features. I don't, by the way.
An Additional Awful thing about Macs and PCs:
Both macs and PCs are dismal when it comes to recovery from crashes.
One the things Iike about unix is that it is hard to completely freeze
the system. The mac is easy to freeze or crash, and so is a DOS machine.
Apple expected programmers to provide the ability to break within a
program and crash recovery IN THEIR PROGRAMS! Of course, most didn't.
Two interesting exceptions are (the original) MacPaint, and HyperCard.
It is nearly impossible to break these programs. Not surprisingly, they
were written by one of the best programmers in the world, Bill Atkinson.
>To print a file, use lpr
Give me a break! Pprinting is only simple in unix when the system adminstrator
has extensively configured the system for the particular computers, programs
and printers being used. You can't simply unplug one printer and plug in
another, and expect it to work. One of the best things about the mac is that
printing is largely device independent. That is, the operating system creates
a generic print file, and universal drivers for particular printers convert them.
This means that old programs immediately work with new printers. It means that
you can buy just about any printer and plug it into a mac, and it works right
away. By Apple Law, the driver must be able to respond to querys by the calling
program about printer resolution,color, etc. Thus, MacDraw prints at the
resolution of the printer connected, be it 72 or 2000dpi. Nothing like this
exists for unix, unless you become a super expert and can tweak down to
the hardware level. This has something to do, in my opinion, why so
many programs on workstations can't save or print their graphics. You
are supposed to take a picture of the screen, since you can't print.
Fair enough. But what if you want to draw an arrow on the picture?
Since you can't save the graphic, you can't do it. This situation is
obviously changing, but the motivation didn't come out of unix
environments, its coming from the microcomputer community.
>I did find that to approximate the power of the
>Sparcstation that I ended up buying, an equivalently-configured MacII would
>have cost MORE.
Does this estimate include a portion of the salary of the system adminstrator,
or the costs of the training courses users should attend?
>There are perfectly good products for all of the
>really popular tasks available for Unix systems. In some cases, the same
>programs are available as on PC's or Mac's (eg. WordPerfect)
Yes, but they cost 2-4 times as much as they do on the mac. Are we really
saving money? Even more important, they weren't invented in unix. Spread-
sheets, desktop publishing, or even fancy word processing were developed
outside of unix, on machines with far less power in hardware and software.
WHY weren't they developed in the more complete, powerful environment of
unix? The fact that unix never generated these types of extremely useful
programs does not bode well for an X-world.
Additional nasty things to say about unix(X)
unix is slow, and X is slower
I've noticed that a lot of unix people actually like the mac--I
think it is because they are used to waiting around. A cold boot of
a unix machine brings back fond memories of my 1984 mac 128 starting
up. And what goes for unix goes double for X--it that 128k look like
a speed demon. This is why many users prefer DOS to unix--things
happen faster, even though the hardware usually is slower.
Is connectivity all that good?
Why does one want to be SO close to everyone. The first worldwide
X virus that comes roaring in over everyone's ethernet will make
it seem less than a good idea. Why do I need the sysop rooting
around in my files? The reason everyone bought personal computers
in the first place (this includes PCs as well as macs, even more
so for PCs) was so that they wouldn't have the sysop telling them
they had to learn things a certain way, or have their computing
witheld! Now we're back to "smart" terminals, and centralized
data storage. Seems like the late 70's doesn't it?
The one place connectivity is fantastic is in mail/ftp type
communication. This is because you use it mostly to talk
to other people, with advantages identical to those provided
by answering machines. Of course it means that nasty little
viruses can wipe out 1000's of machines in a few hours. People
in the Artifical Life group here at UCLA often muse on what
tasty fodder the future worldwide X will provide for their
"living programs" to grow and nuture on.