How Germans write much better code than IBM

THOSE OF YE who follow my articles know that I still keep a lone OS/2 system quietly running - most of the time unnoticed - in my home network setup. It's a dual-PIII system with a gig of ram and IBM's "Warp Server for e-Business", the OS/2-based server OS introduced back in mid-99 by Big Blue and the first Warp version that included Journaling File System (JFS) ported from the corporation's AIX unix to OS/2. I've been running it since I received my first beta copy directly from IBM. And I plan to make this system a living os/2 museum, with plenty of ancient os/2 software to play with, ready to be showed to my grandchildren or to anyone who shows up and makes a contribution to my coffers. Most of the time, however, it lives an uninteresting life, sitting idle waiting for NFS requests from other systems on the LAN.

I recently decided to move the box to a different corner in the South American INQ labs, and at the same time removed its Kensington Expert Mouse -it's a trackball!-, replacing it with a Samsung "Cyber Beetle" PS/2 mouse, another nice affordable (and also several years old) input device with scroll wheel. So, I rebooted hoping for the mouse to be recognized. It wasn't. Fortunately I learned long ago -and the hard way- how to move around OS/2 with just a keyboard.

This opportunity also allowed me to bring the system up to date, including the installation of the latest Firefox 1.0.4, Mozilla Suite 1.7.8 and some other assorted open source software updates (including the latest Java 1.4.2 VM). Not bad for a "dead" OS!, and still a system as useful as I predicted four years ago. Of course I'm not advocating that people migrate towards a platform officially on "life support", yet this proves my point that "perceptions" of an OS's health and popularity have a lot to do much more with the media and broad generalizations than with reality. I doubt anyone is counting my copy of OS/2 as still being used. Not even by IBM.

OS/2 in mid-2005, as useful as any other mainstream OS for web and email...

Still, my lovely Samsung Cyber Beetle wasn't working. And there's nothing strange or overly complicated about mice with a ps/2 bus connection, there's only a handful of protocols that a device can "speak" over the wire, and most of them are widely known by now. So I pointed my Mozilla OS/2 browser to Google, typed "IBM OS/2 mouse driver" and obtained a link to IBM's latest public mouse driver update for OS/2, dated 2004, and available here. After unpacking and installing the driver, the install routine went smoothly. But after the mandatory reboot and as soon as the UI came up and I tried to move the rodent, the pointer moved a bit, only that... it jumped straight to the top or bottom of the screen at the slightest movement, and with one side effect: the system was brought to a halt with each mouse move (even hard disk I/o seemed to be paused).

Back to square one: using the keyboard. I finally Googled around one more time and found the answer to my problem: it turns out that everyone else with a fancy mouse (besides myself) in the OS/2 community -I'm rarely active in OS/2 circles nowadays- seems to be using a third party mouse driver written by some os/2 user in Germany. I installed it in place of IBM's own drivers and now my Cyber Beetle works beautifully. It's called "AMouse" and its home page is over there. This continues a long tradition: OS/2 code (specially drivers) developed outside of IBM is generally of much better quality than IBM's own. Three examples: the best IDE drivers -supporting almost every chipset on the planet and serial ata- are not the ones created by IBM and bundled with the OS but " DANIS506" developed by german techie Daniela Engert, the best OS/2 video drivers didn't came from IBM, but from Scitech SOftware's SNAP Graphics for OS/2 product -licensed by IBM-, and the OS/2 equivalent to "DirectDraw" hardware overlay support on Windows was not developed by IBM back when it had the money and resources to push the OS, but by OS/2 enthusiasts in Ukraine.

However, I know the kind of feedback this article is going to generate "Why are you writing an article about your darn st#pid mouse and an ancient OS like OS/2 that only a {insert favourite expletive here} like you continues to use?". Well, because it highlights some of the recurring problems I continue to see with IBM's software strategy (or lack thereof): #1. arrogance at the technical level ("not invented here" syndrome, which means the company often refuses to acknowledge -until it's too late- that third party technology -or coders- are superior to their own), #2, lack of commitment to software products at the management level (remember the VisualAge series of compilers and IDEs? Remember Viavoice for Linux? what about SmartSuite?. And anyone remembers Hotmedia, the streaming audio and video technology based on Java which could have challenged RealVideo, Windows Media and the like?. All dead or almost-dead products that could still have a market today. And don't make me elaborate on how IBM totally blew the chance of making JFS the default journaling filesystem on linux - to the point that nowadays almost everyone uses ReiserFS or ext3, not JFS).

And this is the same software company which has been making loud requests lately to Sun Microsystems and thinks they could do a better job managing Java than the technology's own creators?. Sheesh!. Maybe I'll take IBM's software claims seriously once again when I don't have to resort to a third party hobby programmer for a quality (read WORKING) mouse driver. Someone please bring Lee Reiswig and his vision back to IBM. µ