Having completed a master's degree in computer science, having used computers of all kinds for many years, having seen the likes of Next, and having used OS/2 for years, I had some great expectations of an operating system. I also knew what was accomplished at Xerox PARC so long ago, and you realize how far behind we are from where we should be. Unfortunately for me, IBM was no longer moving OS/2 forward, so eventually, other things began to catch up capability wise. Where there was once a modest amount of software for the system, the existing companies started to leave. I began to figure out what to do.
I started looking at Linux. I had dabbled in the past, but it had been quite a while. At the same time, I also started checking into Macs since I hadn't followed them much in a while either. Linux was hard, even for me. I kept at it, and I got more familiar with it, and newer distributions kept getting easier. However, it was clear it had a long way to go. It certainly had power, but I was getting to where I was no longer interested in the day-to-day stuff being a challenge, no matter how powerful or slick it was. The Mac was the opposite at the time: no power, but very easy. As much as I liked the Mac, it was still stuck with a Win 3.1 style kernel, and that wasn't going to work for me. I did get some old PowerBooks to play with. I fixed them up, and found they worked well as single application appliances. One just did Quicken - always ready in sleep mode - very good.
Next entry
Monday, May 31, 2004
My Computer
I tried in vain to explain just how awful, in a technical sense, Windows95 was, and where the OS was making things difficult rather than it being the users that just didn't understand. I tried explaining how the Windows Explorer was single-threaded, and this made things feel very like Windows 3.1 at times. I said to see this, copy a sizeable file to a floppy disk using the desktop, and watch the desktop become unresponsive. I would say just right-click on the file and choose "send to", then pick the floppy disk. That's when they would interrupt with, "you can copy stuff to the floppy disk that way? Cool!" I had to admit that the "send to" feature was a handy thing, but I could not convey how limiting the single-threaded explorer was. In OS/2, I could have done multiple drag-drop operations with the same window to re-arrange files without having to wait for the file moves to finish because they went on in the background. The other problem was that DOS/Windows users had been conditioned that copying things to disk was just something you had to wait until it was complete. Even when they had the ability to do other things while copying to a disk in Windows NT, they treated it as a critical operation, and would not task the machine to do anything else in case it "messed up" the copy operation. This explains why they completely accept not using their computer at all while burning a CD. After all, doing anything else might indeed screw things up under Windows, and you were missing the point: the mere fact they were able copy a CD was so amazing that having to do nothing until the process was complete was not a problem to them. This philosophy extended to the computer in general. Windows users also accepted crashes since that had become the norm. No one felt they could do anything about it, but it didn't really matter since everyone else was in the same boat, they could all commiserate and that made things okay.
People like me knew that there wasn't some fundamental law of the universe that meant the computer could only do one thing at a time and that humans had to wait for the machine. What rubbish. The problem seemed to be not that it was technically impossible or hard, but that most consumers didn't know to demand it. At every turn, they would just sigh and bend to the will of the machine. It seemed to be partly because computers were rather complicated, and people just didn't want to bother with all those details to get stuff done more efficiently. Telling them that it was possible to do these things, but you had to use Unix was not an option.
Next entry
People like me knew that there wasn't some fundamental law of the universe that meant the computer could only do one thing at a time and that humans had to wait for the machine. What rubbish. The problem seemed to be not that it was technically impossible or hard, but that most consumers didn't know to demand it. At every turn, they would just sigh and bend to the will of the machine. It seemed to be partly because computers were rather complicated, and people just didn't want to bother with all those details to get stuff done more efficiently. Telling them that it was possible to do these things, but you had to use Unix was not an option.
Next entry
Chicago
The hype of Windows95, then known by its codename of "Chicago", was unreal. By early 1995 it was frightening, and it still would have most of a year before it shipped. It was like being in a horror movie where everyone's body gets replaced or they're under mind control. I had been showing people the wonders of OS/2, and all they would tend to say in response was, "why would I ever need to put a folder inside of another folder?" Nevermind the absolutely mind-boggling other things you could do with folders alone in OS/2, they couldn't get past putting folders within folders. I would try to say it's just like subdirectories in DOS, only you can see them and do all these other amazing things with them that you can't do in DOS. My failure was not realizing that most people didn't even use subdirectories in DOS. The Windows 3.1 Program Manager didn't even let you create subgroups of program icons, let alone let you have icons representing your data files. Sure, they technically had them in the root directory of their C: drive, one for each program they had. However, they did not create any deeper than that. Many people could not think of subfolders to such a degree that they could only think of keeping their data separate by putting it on a separate disk entireley or partitioning their hard drive into several logical drives; all because they couldn't grok subfolders. Little did I know that Microsoft understood this well.
The brilliance of Microsoft is not understanding what their customer wants or needs, but what they are ready to understand. This isn't so much brilliance as just watching which cutting-edge stuff that ordinary people start to get interested in, or by not upgrading their software for so long and listening to what existing customers are shouting as they wave torches and pitchforks outside the window. Soon after I was trying to get someone, anyone, to recognize the wonders of OS/2, even if it was just that you can have folders within folders, I had people telling me about how they were running Windows95, and they could do this neat thing: put folders within folders! When I pointed out that I had shown them this in OS/2 long ago, they just said, "well, but Windows has it now, so it doesn't matter" (is this familiar to you Mac people?). I swear I could hear Rod Serling narrating. I began to search for the pods that grew these zombie replicas of everyone. I later realized that this is just the way it is, and I retreated to my OS/2 world for a while to ponder and wait out the time when Windows ruled the Earth.
I didn't completely retreat, of course. I had to make sure I truly knew the enemy. I tested some of the theories about how Windows95 was constructed. One theory was that even though it was supposed to be a 32-bit, pre-emptive, multitasking, multi-threaded operating system, it was still DOS with Windows 3.1 on its shoulders kicked up a notch with a fancy dress on. The technical bit was that in order to have good backwards compatibility with Windows 3.1, it ran the 16 bit windows system in the kernel with the 32 bit system. Worse, the Win95 kernel was really like the Win3.1 kernel in that it cooperatively multitasked between the 16bit system and the 32bit system. It's just that Win95 was supposed to spend nearly all its time in the 32bit part. What this meant was that when a 16bit Windows program was busy (not all the time), the 32bit system did not execute. I wrote a 16bit Windows program to be busy on demand for several seconds. When this happened, nothing else would respond. The entire user interface would freeze and you could not interrupt it until it was finished. It was later revealed that much of the 32bit APIs just simply called the 16bit ones. This meant that even two "32bit" programs would wind up not multitasking if the both called such functions because the 16bit system was not re-entrant, so only one at a time please. This is why people were told that to be able to take full advantage of Windows95, you had to switch to all 32bit Windows applications (so the 16bit system would never get called), and Microsoft just happened to have a complete set of these ready to go.
Next entry
The brilliance of Microsoft is not understanding what their customer wants or needs, but what they are ready to understand. This isn't so much brilliance as just watching which cutting-edge stuff that ordinary people start to get interested in, or by not upgrading their software for so long and listening to what existing customers are shouting as they wave torches and pitchforks outside the window. Soon after I was trying to get someone, anyone, to recognize the wonders of OS/2, even if it was just that you can have folders within folders, I had people telling me about how they were running Windows95, and they could do this neat thing: put folders within folders! When I pointed out that I had shown them this in OS/2 long ago, they just said, "well, but Windows has it now, so it doesn't matter" (is this familiar to you Mac people?). I swear I could hear Rod Serling narrating. I began to search for the pods that grew these zombie replicas of everyone. I later realized that this is just the way it is, and I retreated to my OS/2 world for a while to ponder and wait out the time when Windows ruled the Earth.
I didn't completely retreat, of course. I had to make sure I truly knew the enemy. I tested some of the theories about how Windows95 was constructed. One theory was that even though it was supposed to be a 32-bit, pre-emptive, multitasking, multi-threaded operating system, it was still DOS with Windows 3.1 on its shoulders kicked up a notch with a fancy dress on. The technical bit was that in order to have good backwards compatibility with Windows 3.1, it ran the 16 bit windows system in the kernel with the 32 bit system. Worse, the Win95 kernel was really like the Win3.1 kernel in that it cooperatively multitasked between the 16bit system and the 32bit system. It's just that Win95 was supposed to spend nearly all its time in the 32bit part. What this meant was that when a 16bit Windows program was busy (not all the time), the 32bit system did not execute. I wrote a 16bit Windows program to be busy on demand for several seconds. When this happened, nothing else would respond. The entire user interface would freeze and you could not interrupt it until it was finished. It was later revealed that much of the 32bit APIs just simply called the 16bit ones. This meant that even two "32bit" programs would wind up not multitasking if the both called such functions because the 16bit system was not re-entrant, so only one at a time please. This is why people were told that to be able to take full advantage of Windows95, you had to switch to all 32bit Windows applications (so the 16bit system would never get called), and Microsoft just happened to have a complete set of these ready to go.
Next entry
Sunday, May 30, 2004
OS/2
I had gone to Dallas to see the launch of OS/2 3.0 and was blown away by the demo. The object-oriented shell was to die for. They showed opening the address book, then dragging a few entries out and onto a document on the desktop causing the document icon to grow a tiny head in its corner to indicate there were some contacts now attached to it. This document was a form letter. Without opening the document, its icon was dragged to the fax printer icon. The system then went into action. It then began to print the document to the fax printer, which caused the word processing program to instantiate the form with data. It recognized that there were address book entries associated, and pulled the fields required (name, address, etc.) from the address book to fill in each version of the letter. Then, the pages could be rendered and sent to the fax printer to be converted for faxing. However, to fax them, you need fax numbers, and the fax printer was able to see the attached address info and pull the fax numbers from it! The fax printer then informed the user that it would have to que up the faxes for sending later as there was no fax modem in the machine. The presenter simply plugged in a PCMCIA fax modem card, the system recognized it, and the fax software began to send out the faxes. Then we got to see how easy it was to use it to connect to the Internet. This was October 1994. Windows95 was a sad beta called "Chicago", and this system could run rings around it. For the next few years, I was in OS hog heaven. I could run DOS programs, I could run Windows programs, and I could even boot real DOS into a virtual machine window. However, you couldn't run some of the newer DOS programs that were using protected mode on their own. The most popular of these was DOOM. It would run, but you couldn't get sound. The stability of OS/2 came at a price. No biggie for me, OS/2 let you boot back into real DOS. However, it was a biggie for many others, and besides, the new Windows Chicago beta was running it with no problems.
Next entry
Next entry
Next
Whilst I was studying computer science, there was a guy in one of my classes who I don't remember much about other than he was a fan of Steve Jobs and what he was doing with Next. He showed me a Next workstation they had in the computer lab and some of what it could do. It was really amazing stuff for 1991, and I ate it up. However, there was that little problem of price again. Next machines were not cheap. I did start to follow what Next was doing. This was a time when operating system technology was a popular topic. There was all sorts of things in the works at various companies. IBM was working on "Workplace OS", which was beat up for never shipping, but it was really some technologies being developed. Windows NT was starting life, and Microsoft was working on what was supposed to be the be-all end-all of systems, codenamed "Cairo", which still hasn't gone anywhere. IBM and Apple began their Taligent effort, Linux was in its formative stages, BeOS popped up later, and I think there was even a small resurgence of AmigaOS as multi-media was all the rage.
The PC people were coming up to the 32-bit operating system decision point since Windows 3.1 was getting old, and there were regular articles talking about alternatives. One I remember was comparing systems to "help" you decide what platform to switch to. It compared Windows 3.1, Windows NT, Windows 95, MacOS and OS/2. The jist of it was that whatever shortcoming that Windows NT had, then Windows 95 could pick up the slack, and whatever shortcoming Windows 95 had, then good old Windows 3.1+DOS could step in. Mac OS and OS/2 were good, but just not enough to beat the Tres Windows Hombres. That was fine as long as you could somehow run all three versions at the same time in the same kernel space.
Next entry
The PC people were coming up to the 32-bit operating system decision point since Windows 3.1 was getting old, and there were regular articles talking about alternatives. One I remember was comparing systems to "help" you decide what platform to switch to. It compared Windows 3.1, Windows NT, Windows 95, MacOS and OS/2. The jist of it was that whatever shortcoming that Windows NT had, then Windows 95 could pick up the slack, and whatever shortcoming Windows 95 had, then good old Windows 3.1+DOS could step in. Mac OS and OS/2 were good, but just not enough to beat the Tres Windows Hombres. That was fine as long as you could somehow run all three versions at the same time in the same kernel space.
Next entry
Comeback
It's not that I really couldn't afford a new monitor, but that it would put a serious dent in the money I was saving to upgrade my computer. By now, I could have afforded a reasonable 386 upgrade, but I had the bug, and I had to have a 486 if I wanted to play computer games reasonably well. That also meant a good sound card. During this time, I was heavily interested in operating systems and was studying them in school. At work, we got a sample disk of IBM OS/2 2.1, and installed it. It was really cool. So cool that I had to run it at home. I found out that a new version was coming out soon that had reduced memory requirements, and so I might be able to just afford the 4MB it would need versus the unthinkable 8MB it currently needed. I began to save even more money, and was able to get the RAM chips off the dead 286 boards piling up turned into SIMMs. I finally was able to get my 486 with 4MB RAM, SVGA (thanks to a used monitor), with a nice sound card and a CD-ROM. I was back. I got my special deal on the new version of OS/2 for $30, and I was having a great time.
Next entry
Next entry
Downgrade
My computer became a new hobby as I now had access to cheap parts and the occasional surplus items, which was good since I didn't have a lot of extra money. I remember getting a tip for a house call I made in the form of a 20MB hard drive and some RAM chips. It turned out that they just happened to be the right chips for a memory board someone had given me a few years before. My computer was getting pumped up with hardware and software, then one day, the 286 chip died. Luckily, we had some spare 286 boards from upgrades, and I was able to use them. Unfortunately, they would not fit in my case, so the board with cards sticking out of it sat beside the case with cables snaking in to the drives. Not pretty, but it worked. Then that board died, and I started to go through all the spare 286 boards there were, the last of which were even slower than where I had started. Then, of course, my nice SVGA monitor died. There was no way I could afford a new one. Luckily, there was a spare monochrome monitor, and my super VGA video card had dual output, so you could hook up the older monitor types and run the card in the old video modes. I was just able to run Windows 3.1 in a low-resolution Hurcules Monochrome mode to get by. My downgrade was a low point.
Next entry
Next entry
Supergeek
My box served me pretty well, and I finished my engineering degree. That Christmas, I started to play with my old Sinclair just for a change. I even learned a few things and wrote a game. I liked games, but I'm just not very good at writing them, so I would to challenge myself. One summer, I wrote two little games in Lotus 1-2-3 just to better learn its macro language. I soon started back to school to get a Masters degree in Computer Science. The aerospace industry in 1989 was not exactly the best job market, and I was now a tremendous computer geek, so it seemed natural to formalize it and be a certified supergeek. Soon, I got a job working at a small local computer store building PCs and LANs and doing custom programming. While I was learning several computer languages at school: Fortran, Pascal, C, assembly, Lisp, Prolog, I was programming at work in Forth, C, Visual Basic and Delphi. All while building computers and hooking up small networks. I was loving it. Once the game DOOM came along, we had some awesome LAN parties even before most people knew you could play the game with multiple players.
Next entry
Next entry
80286
Armed with a bank account with more money than I probably had ever had in it, I went shopping around for a regular PC. This was a time when the clone market was really starting to take off, and it was making it difficult to decide. For some reason, I decided to spend more on the graphics capability - probably after several years of staring at the PCjr's CGA screen. Once I saw a super VGA display, I had to have it. I had beautiful text for several years, but SVGA took quite a while to be fully utilized due to lack of standards, so I had very little good graphics before my hardware was too slow to run it.
I had myself a 286 at a whopping 12MHz (16 MHz just cost too much at the time). I think I did get the full 640KB RAM, and the amazingly vast 1.2MB floppy drive with a 360K drive as well. A 1200 bps modem to connect to the university computer, DOS 3.3, a box of floppy disks, and my tight 12" SVGA monitor with a honkin' 512KB SVGA card. I think I got out of there for around $2400. I remember when writing the check out, I hadn't left quite enough room since I'd not writen a check for such a large amount before.
I left with a Packard Bell made in California (before they became absolute crap) bought from the SoftWarehouse in Dallas, which later became the first CompUSA. Once again, investing in the PC. It didn't even have a hard drive. Why? At the time, I could fit WordStar and the Fortran compiler on one 1.2MB floppy. All my data could fit on one or two disks. The next summer I had saved up for a hard drive as my storage needs increased at a rapid pace. The 40MB harddrive was a vast tract of hyperspace by comparison.
Next entry
I had myself a 286 at a whopping 12MHz (16 MHz just cost too much at the time). I think I did get the full 640KB RAM, and the amazingly vast 1.2MB floppy drive with a 360K drive as well. A 1200 bps modem to connect to the university computer, DOS 3.3, a box of floppy disks, and my tight 12" SVGA monitor with a honkin' 512KB SVGA card. I think I got out of there for around $2400. I remember when writing the check out, I hadn't left quite enough room since I'd not writen a check for such a large amount before.
I left with a Packard Bell made in California (before they became absolute crap) bought from the SoftWarehouse in Dallas, which later became the first CompUSA. Once again, investing in the PC. It didn't even have a hard drive. Why? At the time, I could fit WordStar and the Fortran compiler on one 1.2MB floppy. All my data could fit on one or two disks. The next summer I had saved up for a hard drive as my storage needs increased at a rapid pace. The 40MB harddrive was a vast tract of hyperspace by comparison.
Next entry
Peanut
In 1985, I was given a PCjr as a graduation gift to help me through college. This was primarily because it was much more affordable than an Apple II, Mac or a regular PC. For the first year of college, I still used my trusty Sinclair quite a bit since I could do magic with it and hadn't yet gotten comfortable with the new level of complexity that the PC was. I remember realizing that I almost completely knew my Sinclair (I would realize later that there was still quite a bit I didn't knwo). The PC was clearly far more involved; at least in the hardware and operating system. Initially, all I cared about was that the BASIC was quite lacking, although it did have more capability, but that was just because the PC had more power than my Sinclair.
It took me a while to realize the problem of owning a PCjr. It didn't really matter for a while until my needs grew. By 1988, I was needing more to finish my final years in Aerospace Engineering. In the summer of 1988, I wrote a BASIC program on my PCjr to make the summer job I had much easier. Easy to do when you've got lots of spare time when you're young. At some point, I just took my whole system to work, set it up at my desk and began keying in new data as it came in and entering existing data when I had nothing else to do. It worked great. So great that I was able to sell the program to the company along with my entire computer system before I went back to school. I used the money to help buy a new computer - a regular PC. I had managed to get out from under a PCjr for not much less than it cost and then some for the software, which they ran for about another 10 years. Before that, I was certain that I'd be stuck with the PCjr lemon, unable to get rid of it. I just had to find the right job for it.
Next entry
It took me a while to realize the problem of owning a PCjr. It didn't really matter for a while until my needs grew. By 1988, I was needing more to finish my final years in Aerospace Engineering. In the summer of 1988, I wrote a BASIC program on my PCjr to make the summer job I had much easier. Easy to do when you've got lots of spare time when you're young. At some point, I just took my whole system to work, set it up at my desk and began keying in new data as it came in and entering existing data when I had nothing else to do. It worked great. So great that I was able to sell the program to the company along with my entire computer system before I went back to school. I used the money to help buy a new computer - a regular PC. I had managed to get out from under a PCjr for not much less than it cost and then some for the software, which they ran for about another 10 years. Before that, I was certain that I'd be stuck with the PCjr lemon, unable to get rid of it. I just had to find the right job for it.
Next entry
A long time ago, etc.
I remember back in 1983 going to a computer show at the Info Mart in Dallas. One of the things I saw there was an Apple Lisa. I only remember that I didn't spend much time with it and that I didn't fully understand what I was seeing. I do remember noting two things: It was a bit pricey at $10000, and that it looked good to me because it had a black on white display like my Sinclair.
Somewhere in 1984 to 1985, I remember checking out the early Macs and getting a very mixed first impression. I liked the system, but it did tend to crash a lot - how much of that was due to beat-up demo units, I don't know. I was looking mainly because I had been reading all this hoopla in the computer press about how great "pull down menus" were until I was sick of hearing about it. To vent my frustration, I programmed up some pull-down menus for a program on my Sinclair. A little man would come running from the side of the screen, jump up and pull down on a cord, thus pulling down the menu (rather slowly animated), then wait until you made a choice and let it roll back up as he ran off the screen. Once I saw the Macs in the flesh, I better understood the usefulness, and promptly went home to redo the menus and entire screen on my program, which ended up looking like a clone of the Mac screen with the same pull-down menus (sans little helper man) that even blinked the choice you made before vanishing. Of course, my program wasn't an entire operating system, but it certainly was easy to use.
Next entry
Somewhere in 1984 to 1985, I remember checking out the early Macs and getting a very mixed first impression. I liked the system, but it did tend to crash a lot - how much of that was due to beat-up demo units, I don't know. I was looking mainly because I had been reading all this hoopla in the computer press about how great "pull down menus" were until I was sick of hearing about it. To vent my frustration, I programmed up some pull-down menus for a program on my Sinclair. A little man would come running from the side of the screen, jump up and pull down on a cord, thus pulling down the menu (rather slowly animated), then wait until you made a choice and let it roll back up as he ran off the screen. Once I saw the Macs in the flesh, I better understood the usefulness, and promptly went home to redo the menus and entire screen on my program, which ended up looking like a clone of the Mac screen with the same pull-down menus (sans little helper man) that even blinked the choice you made before vanishing. Of course, my program wasn't an entire operating system, but it certainly was easy to use.
Next entry
Saturday, May 29, 2004
Take the White Pill
This is about my experience switching to a Mac. Hopefully, any tips and solutions I come across will be useful to others. The difference from most "switcher" stories for me is that this switch has taken 20 years. Well, to be fair, maybe not quite that long. A few years ago, I aquired several old PowerBooks and a Mac SE/30 for a couple hundred dollars to play around with, but they were not used as primary machines. I'll start with a quick history to catch up and get ready for the arrival of my 12" iBook G4. Now that I've taken the White Pill, we'll see what happens.
Next entry
Next entry
Subscribe to:
Posts (Atom)