Oopsie

Back when I started exploring the Intel Mac, I explored Javascript implementations, and concluded that it was better on the Intel iMac than on my desk-lamp (G4) iMac. Well, yesterday, I hit one too many beachballs in Safari on the G4, and downloaded the latest version of Firefox. To my amazement, the numbers in the Web Sudoku squares behaved appropriately, leading to the suspicion that the problem was Safari, not the platform it was running on. To test this suspicion, I fired up Safari on the Intel iMac, and discovered that it, too, flubbed the Sudoku display. There's a lesson here...

Onward into the Abyss

One of our researchers came by to consult me about equipment purchases for a grant whose funding begins on April 1. A new grant is always cause for celebration, but when computer hardware is in a state of flux, it's hard to know what to advise researchers. She really wants to order a new MacBook for her research assistant, based on the built-in iSight. Her research concerns signed languages, so video-conferencing is very useful in collaborations. However, a crucial aspect of the research involves annotation of video files, using ELAN, a program developed at the Max Planck Institute for Psycholinguistics in Nijmegen.

The MPI web page doesn't indicate whether or not the program runs on an Intel Mac, merely that it requires OSX (or Windows or Linux), without specifying any specific hardware requirements. And the latest (bug-fix) release was in November 2005. Furthermore, the web site makes heavy reference to the Java Runtime Environment. But, nothing ventured, nothing gained. I downloaded and installed the program, scrounged a movie file (though the preferred format is MPEG-1, and despite the fact that the tutorial's author had difficulty with .mov files, albeit in a Windows environment). It was a little difficult to figure out how to do anything with the program, but it definitely runs, and I was able to insert a few annotations, and use the interface to scroll around my movie.

More on Matlab (the good, the bad, and the ugly)

The good news is that Mathworks has issued a new version of Matlab.

The ugly news is that it's Release 2006a, which is also described as Matlab 7.2. Yes, Mathworks has decided to confuse us with yet another numbering scheme for Matlab, just when I'd finally internalized that R14 and version 7 were the same thing, with R14 Service Pack 3 corresponding to version 7.1 (as distinct from 7.01, which was R14 Service Pack 1).

The bad news is that the Macintosh system requirements specify that this latest version requires a PowerPC G4 or a PowerPC G5 processor. In other words, it won't run on a Core Duo. The press release about the change in name indicates that the next scheduled update is in September 2006. There is no information anywhere on the website that I've been able to find concerning plans for MacIntel version of the program. Postings on comp.soft-sys.matlab by Mathworks employees are understandably vague, but indicate an intention to provide a compatible version. It's these employees who suggest that Matlab will run on MacIntel, but only from the command line, with the flag -nojvm (that is, without the Java Virtual Machine—that's where the problem apparently is). However, at least some newsgroup posters have had problems when attempting to do so.

So, for now and for the foreseeable future, lack of Matlab could be a deal-breaker for scientists thinking about upgrading to a new Mac.

Coming to grips with the demise of Classic mode

I was doing some trouble-shooting yesterday about difficulties one of our researchers was having printing from Classic mode. In that case, opening the Chooser was sufficient to prod the printing sub-system into working again. But that got me to thinking. In the Mac world as a whole, there have always been killer apps that either didn't get issued for new hardware, or whose delay in upgrading kept users with outdated hardware. There are programs, such as Adobe Premiere, that never came out in OSX versions, perhaps in part because of Apple's own video editing software, Final Cut Pro.

Of course, OSX had a built-in emulator for the pre-OSX environment, and many of our niche programs run just fine in that, though there are sometimes quirks regarding printing or sound-playback. This hasn't been a problem. But it will be. The Intel-based OSX has no Classic mode. This means that, as our users upgrade, they'll have to adjust.

Here are some programs that will cause problems.

One researcher has a huge amount of class-demonstrations written in True Basic. A FAQ on the True Basic web site, which seems not to have been updated in several years, promises that an OSX release of True Basic is forthcoming. But, I'm not holding my breath. Now, there are other Basic programming environments (and for what this is used for, there's no reason to send her off to learn some variant of C+ or for her to rework all of her examples in Matlab), but she's got her library in this one. So that's a problem.

Even though the statistics programs SuperAnova and StatView haven't been supported, or even sold, for years, they still run just fine in Classic mode. Granted, SPSS, which runs in OSX (including on Intel-based machines) is far more powerful, it, not surprisingly, has a much steeper learning curve. There's a niche for a basic, menu-driven statistics package, so I'm going to have to renew my search.

We also have a few users with legacy files in SoundScope, a sound recording and analysis program. This program runs fine in Classic mode, but their web page, copyright in 2005, simply states that the program does NOT run in OSX native mode. These folks are also using Praat and, I hope, will be receptive to switching over. I think the only thing SoundScope will do that Praat won't do is produce an on-screen oscilliscope display.

All is not lost, however, even if replacements for these (and other) programs can't be found.

Enterprising sorts have worked out an open source Classic emulation mode called SheepSaver that will run on the Intel Macs. Because it comes from Unix-land, installation is anything but straightforward, and involves, among other things, locating a ROM file from a Mac that will run OS9 (or earlier) and a valid pre-OSX System Folder. Looking at the documentation and help files for this, however, makes it clear that installing and running this environment is a non-trivial task, and it might not be a worthwhile use of time and effort to make it work. But it is a possibility.