An Apple in Academia

It’s all Waterloo-Maple‘s fault, really: if they had maintained a version of their computer algebra system for the Amiga, I wouldn’t have found it necessary to switch to Mac. Or maybe it’s Commodore’s fault for mismanaging themselves into oblivion; I don’t know. Either way, I became painfully aware three years ago that my little Amiga would no longer satisfy my computing needs. I needed a new home computer.

The Problem

My biggest need was research. I didn’t have the time or energy to program a complete symbolic computation package for the Amiga; besides, I wasn’t supposed to program a symbolic computation package; I was supposed to use a symbolic computation package to research mathematics. There had been a version of Maple for the Amiga back in the early 90s, but by 2001 it was hard to come by. Waterloo-Maple didn’t even reply to my email inquiring about it. The Amiga, alas, had to go.

Previous attempts at a solution

I’d had experience with Microsoft software, having once owned a Òbeige boxÓ back in the days of Windows 3.1, and of course it’s nearly impossible to go through the world of computers without at least once using a Windows machine. My Programming 101 was done on MS-DOS machines that ran Borland’s Pascal. Those frustrating experiences with MS-DOS and Windows drove me into the arms of Amiga: the first machine I ever used that, with its combination of pre-emptive multitasking, color GUI, and integrated command-line interface (unique on home computers of the time) felt neither like a straitjacket nor like an unmanageable nightmare. It was on the Amiga that I really sunk my teeth into programming, both in C and in Modula-2. It was with TDI’s and later Benchmark’s Modula-2 compilers that I wrote my first symbolic computation software: to handle arbitrary-precision integers, to manipulate continued fractions, and to draw the graphs of particular solutions to two-dimensional differential equations.

In the years since embracing Amiga, I had seen nothing from Microsoft to stimulate my enthusiasm for the Windows platform. So, Microsoft was off the table. In 2001, that left Linux and Apple.

My other criterion was that my new computer be portable: two months before, I had been visiting my parents’ house. Driven by the desire to crack a problem, I stayed up all night reducing S-polynomials by hand. There is a reason we have computers, my friends: as I re-discovered the hard way, that reason is to make mathematics easier.

Why not Linux?

A few months prior, the symbolic computation professors at my university had purchased workstations for research purposes. This being a university, we decide things by general consensus — i.e., the loudest voices win 🙂 The consensus was that, since (a) Linux was free, (b) the university offered a modified RedHat distribution, and (c) two of the grad students ran Linux boxes at home, we should standardize on Linux. At the time, Linux really did seem to be the wave of the future.

One professor dissented, and purchased three dual PowerMac G4s for his students. The professor most in favor of Linux chose what I consider to be a most curious argument to express his distaste for the Macs: they made us look “less professional, like a mathematics lab instead of a computer science lab.” The fact that we computer algebra students are a lab in a mathematics department, and not a computer science department, didn’t seem to impress him.

We did not purchase the machines all at once, but over a period of several months. The first two purchases went very well. Mine did not: at the time, X didn’t support the Matrox Millenium G450 video card — or at least, the X server provided by our university’s version of Linux didn’t. Realm Linux is closely tied to RedHat Linux; at the time I tried to install it, it was built from RedHat 6.something; today they build off Fedora.

I was a complete newbie to this; in more than ten years of using Amigas, I’d never had this problem (probably because I was lucky). I couldn’t figure out how to make X and the G450 play nice, and there were working machines nearby, so — I worked on those other machines. On occasion, I would try to figure the thing out, but to no avail. The two colleagues who had made the successful purchases, and who ran Linux workstations at home, were unhelpful. A third officemate finally took pity and sat down with it one morning. Late that afternoon, he figured out how to make the video card work with the X server, and Žowyn.math.ncsu.edu was finally online. (After graduation, this gentleman rather appropriately took a job with RedHat, testing the installation of the operating system on differently-configured systems.)

My advisor purchased a machine nearly identical to mine, but his difficulties were even worse. He gave up on Linux altogether and installed Windows on it. There were serious problems with that, too, but I’m not one to beat the dead horse of the problems with Microsoft’s default security settings.

My new machine

Given that experience, buying an Apple was a no-brainer. I bought a brand-new 500MHz iBook, with all of 256MB RAM. (This, compared to the Amiga’s 33MB I had never been able to fill in ordinary usage.) I really wanted a blueberry clamshell — if you’re going to go crazy, go crazy in style, no? — alas, those had been phased out. I was stuck with a colorless, all-white iBook. Yuck. At least it ran Maple. I named her Parvula, and hoped for the best.

This was August of 2001, and OS X was still new (and still slow as molasses in winter). I didn’t mind the slowness too much; what mattered to me was that it ran Maple, and it wasn’t Microsoft or Linux. My machine dual-booted OS X and OS 9. Maple ran only in “classic” mode: that is, it required OS 9 emulation within OS X. So starting Maple was a chore that required me to wait, and wait, and wait… wishing all the time that my university had standardized on Mathematica instead, with its polished Aqua interface. To my dismay, the CD-burner didn’t write under OS X 10.0, but that was fixed before too long.

At the time, there were very, very few Apples around the department. Linux was all the rage. There was talk that our university would soon be taking the cost-cutting measure of eliminating Windows and moving everything to Linux. That long-promised switch to Linux never actually took place, although one public lab in my building has switched to Linux and OpenOffice, another to Linux and CrossOver Office.

Insane, perhaps, but not entirely isolated

Most of my colleagues thought me insane, and said so.

Ray was the only one who owned an Apple. (Note: names have been changed, to protect the guilty.) He owned a nice PowerBook, and was lusting after a Cube. (Remember those?) He and I would chatter about our machines while my officemates rolled their eyes, or asked when we would remove that Mac virus, and install YellowDog Linux. Both of us were interested in the huge amount of Unix software being ported to OS X, through the outstanding work of the Fink project. I invested some hours learning how to fink -install some-program and later to dselect. Then came Fink Commander, and life became easier. Coming from a Realm Linux world where state-of-the-art was the antiquated rpm –install some.rpm, then gnashing my teeth in dependency hell (I never had to worry about that during ten years of installing Amiga software — but maybe I was lucky), it seemed just a little ironic that Debian’s dselect masterpiece was perfected… on a Mac.

Soon thereafter I was using my iBook not only to run Maple computations, but to type results and reports using Lyx, XFig, and the whole gamut of TeX-related software, and showing Ray how to get fink, X, and the rest running on his PowerBook. (Ray however prefers TeXShop to Lyx.) I’ve even downloaded Loria’s SmartEiffel browser and refreshed my memory of the Eiffel programming language. Both of us also enjoyed indulging in our non work-related, “consumer” uses of our Apples; here Ray had a definite advantage, since with a video camera he could create an entertaining video of his two year-old son running buck nekkid about the house).

We were both pretty happy with our machines.

Fast-forward nearly three years

Realm Linux finally became useable. With the advent of KDE 3.1, and the attendant eye-candy from Everaldo’s Crystal icons and Keramik, Linux also became pleasant to use. (Except when RealPlayer decided not to work — but Žowyn is a work computer, and there are more important concerns.) For a while, I explored the idea of converting the iBook to Yellow Dog Linux. In the end, I decided to stick with OS X…

…Mostly because OS X has also spent the last three years improving. I don’t have to pay full price for the OS upgrades — nowhere near full price, to tell the truth — so I didn’t mind the yearly subsidy of Apple’s software group. Besides, the upgrades have been worth it. Upgrading has been remarkably smooth overall. It is especially gratifying to type this at a time when one of my colleagues has struggled for a full week to get two of the Linux machines in the office to upgrade from Realm Linux/RedHat 8 to Realm Linux/Fedora Core 1.

Maple 9 finally runs natively on the Mac, albeit only in the new, ugly, slowish, Java-interface worksheet mode. Still, it allowed me finally to eliminate Classic emulation from my machine. In addition, I ran it overnight and found a counterexample to a conjecture I’d been trying for a long time to prove — no wonder I couldn’t prove it; it was false! With that one stroke, my PowerPC G3 machine justified my investment. The dual 1GHz Pentium running Linux SMP didn’t find the counterexample; such are the vagaries of Fortune and pseudo-random code. 🙂

The iBook has proven itself enormously useful in pretty much every aspect of my professional life, as well as many aspects of my non-professional life. The most mundane task of all is record-keeping for the occasional class I teach: recording grades and attendance (Appleworks), writing syllabi (Mozilla), tests (TeX), etc. There’s also a lot of fun to be had in digital photos, iTunes, and reading OSNews with the Camino web browser.

A simple twist of fate…

The distribution of Linux boxes in professor’s offices has changed; many professors have purchased Linux boxes, but some have replaced their Linux boxes of three years ago with Apples. I can’t tell you why; I don’t know these professors, and I’m a bit too shy to walk into a professor’s office and ask, “Say, I’ve noticed the old RedHat workstation has been replaced with a PowerMac — what’s up with that?” But there have been a few changes.

My roommate, who works in applied mathematics, decided to buy a G4 iBook, due both to my positive experience with it, and to his own advisor’s use of Apple hardware for his research. They run Matlab mostly, occasionally Maple. He has been quite satisfied with the iBook. I found myself occasionally on call to help him get software running on it. The most pernicious problems was getting his wireless ethernet router to communicate with his iBook, but that was due to the router’s shoddily-written manual, which gave the wrong name for the wireless network.

The biggest surprise came some weeks ago, when one of my office’s biggest Linux advocates showed up with a 12″ PowerBook. Peter had just won a research award, and wanted a portable to take his work with him to conferences, or even on the road. He decided to buy a PowerBook. Why? He didn’t want to spend two or three weeks researching which laptops were supported by Linux; nor did he want to waste time getting CygWin to work on a Windows box. He simply ordered a PowerBook, got it in the mail, and was using it the following weekend at ECCAD 2004 to give a talk using his TeX-generated slides (works great with Preview — and probably with Acrobat as well) and to demonstrate the Maple program he wrote to implement his research. He won’t shut up about how cool iTunes is.

This caused quite the stir; Peter brought his PowerBook to a party at another colleague’s home. The goal was to test whether it worked with the apartment’s free wireless internet access. It seems Walker (the colleague hosting the party) has been considering buying an iBook himself, but he wanted to make sure the Airport Extreme card would pick up the apartment’s network. It was as easy as opening up the PowerBook: it woke from sleep, detected the network, and popped open a dialog, naming the network and asking if Peter wanted to join it.

The other mathematicians at the party (and two computer scientists as well) were eager to set their eyes on Peter’s PowerBook; it made the rounds and won general acclaim. Quite a few expressed favorable opinions, and said they would look into acquiring one. There were the usual comments on its stylish looks and its gorgeous interface, including from Peter — this from the man who used to mock me for my iBook’s “eye candy” 🙂

Triumph of triumphs: Peter’s advisor — he of the opinion that we Computer Algebraists should be using Linux instead of a Mac — was impressed at the conference by Peter’s PowerBook, and said he might have to get one.

Apple is making inroads in academia 🙂 My colleagues are catching up with my wisdom 😉

But will it mean anything?

I don’t know if Apple is aware of the potential market it has in academia. The Virginia Tech supercomputer cluster aside, I haven’t seen much of a marketing push at the university level. — Sure, there’s been a general push to attract users of Unix, but as someone who works in academia, I don’t feel Apple is making anywhere near the effort to attract me that they could. People like me are precisely whom Apple should be targeting: scientists and mathematicians who use Unix software to solve problems, and who can afford the price premium (real or imagined) necessary to avoid the troubles inherent to Linux or Windows or Sun or… Admittedly, I may not run in the right circles to notice such a push.

Universities have a legacy of using Unix; mathematicians and computer scientists are familiar with it, as are most natural scientists, and most of our software runs already on some flavor of Unix. Porting legacy Unix software to OS X is relatively painless (especially when compared to porting it to Windows); the Fink project has already made much progress in this regard. OS X is much, much easier to use than traditional Unix boxes: all the power, little of the pain. I don’t know what the statistics are in terms of installed computers, marketshare, &c.; to tell the truth, I don’t care. What I care about, is that I get my work done, using the software I need, in an environment that is user-friendly, and Òjust works.Ó

Okay, so what exactly is your point?

A month or two ago, another OSNews author asked the question, can you get “serious work” done on a Mac? His answer: maybe other people could, but for various reasons, he couldn’t. As I recall, his problem was getting his Apple to work in the Windows-dominated environment where he worked; at a low point in the article, he pointed out that Microsoft Office for OS X wouldn’t properly read PowerPoint files written by Microsoft Office for Windows.

For obvious reasons, I was genuinely dismayed to read the article. I’m lucky enough to work in an environment where the IT department is half-competent and runs a network that is truly open. All our information is shared on a centralized network that runs on AFS. I hadn’t invested huge amounts of my time and data using programs that don’t play well with other, like Microsoft Office. It isn’t hard to connect my iBook to the network, and the common Unix foundation underlying Linux and OS X make it easy to control Žowyn from Parvula. Getting the two machines to cooperate has been a snap.

Based on my experience: Yes, you can real get work done on a Mac, and a growing number of academics is doing just that. But has Apple noticed, and will they capitalize?

About the author:
Jack Perry is a research assistant at North Carolina State University. For some reason he can’t fathom, one of his web pages is (at the time of this writing) the #1 Google hit when searching for S-Polynomial reduction. (Something more useful ought to be #1.) In his spare time, he laments the greater popularity of the curly-brace languages over languages where statements such as

p++ = --q ? c = a(*p, **q) != *b(&p, q) : c = a(*p, **q) == *b(&,q);

are not merely “a bit much”, but make no sense at all 🙂

52 Comments

  1. 2004-06-03 6:08 pm
  2. 2004-06-03 6:11 pm
  3. 2004-06-03 6:16 pm
  4. 2004-06-03 6:37 pm
  5. 2004-06-03 7:35 pm
  6. 2004-06-03 8:14 pm
  7. 2004-06-03 8:18 pm
  8. 2004-06-03 8:21 pm
  9. 2004-06-03 8:46 pm
  10. 2004-06-03 8:56 pm
  11. 2004-06-03 9:11 pm
  12. 2004-06-03 9:13 pm
  13. 2004-06-03 9:29 pm
  14. 2004-06-03 9:37 pm
  15. 2004-06-03 9:37 pm
  16. 2004-06-03 9:37 pm
  17. 2004-06-03 9:46 pm
  18. 2004-06-03 9:50 pm
  19. 2004-06-03 9:59 pm
  20. 2004-06-03 10:51 pm
  21. 2004-06-03 10:54 pm
  22. 2004-06-03 11:22 pm
  23. 2004-06-03 11:28 pm
  24. 2004-06-03 11:43 pm
  25. 2004-06-03 11:50 pm
  26. 2004-06-04 12:12 am
  27. 2004-06-04 12:26 am
  28. 2004-06-04 12:29 am
  29. 2004-06-04 12:34 am
  30. 2004-06-04 5:06 am
  31. 2004-06-04 5:31 am
  32. 2004-06-04 7:36 am
  33. 2004-06-04 7:46 am
  34. 2004-06-04 8:09 am
  35. 2004-06-04 10:25 am
  36. 2004-06-04 12:59 pm
  37. 2004-06-04 1:01 pm
  38. 2004-06-04 1:12 pm
  39. 2004-06-04 1:16 pm
  40. 2004-06-04 1:28 pm
  41. 2004-06-04 1:53 pm
  42. 2004-06-04 3:50 pm
  43. 2004-06-04 3:54 pm
  44. 2004-06-04 7:36 pm
  45. 2004-06-04 7:59 pm
  46. 2004-06-05 3:39 am
  47. 2004-06-05 5:56 pm
  48. 2004-06-05 6:11 pm
  49. 2004-06-05 10:26 pm
  50. 2004-06-06 4:01 am
  51. 2004-06-06 5:09 am
  52. 2004-06-06 10:55 am