BIOS (LIFE), The Rest Is History.
The development and evolution of the computing industry,
from a "Geek's" viewpoint.
Back in 2006, I decided to write an article on how may of us "cut our teeth" in computing. This article reflects the development and evolution in the computer industry from my personal expierence.
The development and evolution of the computing industry,
from a "Geek's" viewpoint.
Back in 2006, I decided to write an article on how may of us "cut our teeth" in computing. This article reflects the development and evolution in the computer industry from my personal expierence.
BIOS, meaning life, power released , life that has sprung from our Creator's breath to bring life to man. Who created this BIOS, or "Basic Input Output System", man bringing life to a machine, a
machine created by man for his purposes. The existence of the personal computer seems to have become a part of our everyday life, intertwined with our day to day activities. Think about it, even if we choose not to own one, there is a computer involved somewhere, linking our choices to an outcome that has somehow been processed, analyzed, and crunched in a silicon brain. When the founder of digital Research, Dr. Gary Kildall wrote a primitive operation system called CP/M or Control Program for Microcomputers, it was designed to work only with IBM compatible soft sectored floppy disks. As the need arose to use CP/M with non IBM compatible floppy disks, Gary isolated the code that integrated the format of the disks to the rest of the operating system into a separate "BIOS" module that could be easily change to meet the requirements of other disk formats. The name stuck, the rest is history. This BIOS has evolved to be the life blood of our PC today. It seems to keep the hardwired components of our modern day PC working together in harmony. Does this scenario sound familiar? Man always strives mimic what God has created.
In 1987-88, while oblivious to any other ground breaking developments in technology, I was happily delving into the innards and secrets of my Commodore 64. After all, the secrets it held were captivating to no end! Before that I tinkered with an APF M1000. It sported a Motorola 6800 microprocessor, built in basic interpreter, and 8 kilobytes of system RAM (Random Access Memory) which I soon modified to 24 kilobytes! It wasn't until I took a home study course in computer technology, that I was introduced to the IBM clones. As part of the course, the lucky student built a brand new shinny Heath/Zenith clone PC complete with DOS 2.0! After it was all said and done, what was I to do with a computer with a monochrome screen, and only 128 kilobytes of memory? At that time memory was expensive (to me anyway). So shelling out the cash to bring the memory up to a whopping 512 kilobytes gave me no satisfaction. Then I found a whole new world of PC technology beyond the excitement that my Commodore 64 could provide to captivate my attention, and have been hooked ever since.
Early on, there were three main driving forces that shaped the IBM compatible PC industry, IBM, Intel and Microsoft. Of course there were plenty others before and after, who contributed immensely, and they are still contributing to the industry today. But, together IBM, Intel and Microsoft linked the ground working tools that the industry first embraced on a large scale to get the ball rolling. Microsoft DOS was actually a clone of CP/M. When IBM's interest in using the CP/M operating system in their first personal computer was put off by Gary Kildall, they turned to Microsoft for an operating system for the computer that would change the way we live. William Gates at Microsoft saw an opportunity waiting, and agreed to provide an operating system for IBM's new child. Microsoft having no solution of it's own, turned to Seattle Computer Products. Early in 1980, Tim Paterson, who worked for Seattle Computer products, had written a cloned version of CP/M called 86-DOS to run on Seattle's S-100 bus 8086 based microcomputer. Up until that time, CP/M had not been ported to the Intel 8086 microprocessor platform. So Microsoft licensed 86-DOS from Seattle Computer and presented it to IBM as Microsoft Disk Operating System 1.0. Since IBM could not find any other suitable offerings, they agreed to strike a deal with Microsoft to use MS-DOS in their new 8086 based PC. So in July 1981, Microsoft purchased all rights to 86-DOS, and in August 1981, IBM announced the IBM PC Model 5150, which became available for purchase in the fall of 1981. The PC business was a good run for IBM, 23 years. In the fall of 2004 IBM announced it would sell it's PC division to a Chinese manufacture, Lenovo.
If you were to ask who is responsible for this invention that has change our lives, I could only answer 'no one in particular.' It were those with vision that built upon, and combined the work of others to create what we have today. How far back shall we look for the roots of the PC? It is what we experience that we remember the most. We program software to run on a specific computer platform. But we are programmed by experience. It seems that we were created to learn from our past mistakes and accomplishments, and desires. Shall we remember Dr. Ted Hoff, who in 1971 created the Intel 4004, which was the first commercially available microprocessor. I remember the wonder in reading about the Altair 8800 computer in Popular Electronics. If you had enough money to spare, you could actually buy it in kit form. Two college students, William Gates, and Paul Allen wrote the BASIC language interpreter for the Altair 8800, and later formed Microsoft to sell their software product. And let's not forget about Steve Wozniak and Steve Jobs who built the Apple computer.
It wasn't until I pursued my first home study computer course in 1980 that I entered into the world of digital logic circuits. Similar to the Altair 8800, the finished product of the course that I took was a good sized box with many switches and lights with which my new toy was programmed. Computers then were in their most basic form, 0's and 1's, digital logic. How many of us every day ignore the fact that turning on the light switch in our homes is a digital function, off is '0', on is '1'. It is the same as our PC's today do it, except they manipulate the 0's and 1's elegantly on a far greater scale. In high school I created a electrical tic-tac-toe game. It was driven by relays, dozens of them. It was hardwired logic, and un-beatable as long as it was allowed to make the first move. I was only one person whose creation only served myself. But what about the countless others that have made small contributions similar to my logic driven invention, which were used to build upon, to create what we have today. It is said that ENIAC, the first vacuum tube computer was contained in a room 30x50 feet, weighed 30 tons, and contained 17,468 vacuum tubes. It could executed 5,000 additions and 300 multiplications per second. In contrast the Intel 4004, an early microprocessor contained 2300 transistors and executed 60,000 instructions per second.
How fast shall we go? Steve Gibson, founder of Gibson Research, wrote a column for Infoworld magazine for a number of years. In 1989, he wrote something that no one including himself, ever would have thought could happen. He wrote, " We've grown so accustomed to the clock rate ride that it's easy to simply anticipate that the joy-ride will continue through coming years with announcements of 40, 45, 50, 65, 75, 100, 150 megahertz systems. It won't happen. You're sons will not be bouncing their grandsons upon their knees telling them of the days when PC's didn't all run at 12 gigahertz. Just as the performance of dynamic RAM memory bottomed out several years ago, we've come to the end of our years of clock rate joy riding... completely regardless of how much we're willing to pay. " I imagine he is chuckling about that statement today. Even though the industry has found ways to speed up microprocessors at lower clock speeds, we are well into the 4 gigahertz range. How long until we hit that 12 gigahertz mark?
I started my venture into the PC word with a Heath/Zenith PC-XT 8088 8 megahertz clone when the Intel 80386 CPU was arriving on the scene. Heath/Zenith took a different approach to design compared to the standard today. Instead of a main board, they designed a backplane board which all the other boards plugged into. As I recall, the CPU, and memory sat on a single board, The I/O (Input Output) functions on a separate board, and also the video on a separate board. The 360 kilobyte 5 1/4" floppy drive held more than that of my Commodore 64, but I had and no sound, a green monochrome screen, and no hard drive. How did I ever survive that? Computer parts were quite expensive at that time, so I was always a few steps behind on the technology. This gave me the motivation to learn how to squeeze every last ounce of speed out of my hardware. If I were to teach a class or give a lecture on using computers, I would most likely find it difficult, unless I were permitted to explain the bowels of these perplexing machines. It seems that today some of the fun has diminished because now we are able to quickly boot up our PC's, and open an abundance of graphical windows while also doing a multitude of other things (Multitasking), most of which is not prominent or readily noticeable to us. We have plenty of RAM, plenty of hard disk space, plenty of resources. This is good news for the general population of computer users because it relinquishes more time to focus on the task at hand. But also bad for inquisitive hardware creatures like myself just because things are running just to darn smooth lately! In those days (not that long ago) I envied anyone who could afford to keep up with the technology with all the latest computer gadgets. In my circle of computer enthusiasts the driving force was how can we get more while spending less! I soon discovered that unlike my trusty Commodore 64, this inscrutable PC seemed to have unlimited potential to expand. So there I was, looking at empty sockets, how tempting it was to easily increase the RAM! After assembly, my PC-XT clone which only sported 128 kilobytes of RAM. To increase it in any amount required a set of nine identical DRAM (Dynamic Random Access Memory) chips. Unlike today's memory module, each chip was flat with downward pins on each side which fit nicely into a socket. Having populated my sockets with nine more 128 kilobyte, and nine more 256 kilobyte chips, I now could enjoy 512 kilobytes of system RAM. I suspect Dr. Robert H. Dennard, the inventor of DRAM may have not realized the extent of his creation. Working for IBM at the time, he had an idea to design a simple memory cell using a single transistor and capacitor, a device similar to a short term storage battery. His memory cells, could be accessed or written to independently of each other. Remember those 0's and 1's? Each cell in a bank of memory cells can be changed to '0' - off, or '1' - on. For many years I have worked in the Engineering Department of a local hospital. In doing so I have been given the opportunity to see first hand much of computer history, including Dr. Dennard's invention in action. Presently we have many ways to keep in touch and keep track of each other. But in the past, the telephone switchboard consisted of rows of jacks and plugs to manually connect calls to their destination. In those days, thanks to Dr. Dennard's invention, when entering the building our MD's entered a 3 digit code into a key pad while pressing either an IN or OUT button. This simply programed one of Dr. Dennard's memory cells either on or off. Accordingly an indicator light referenced the state of the memory cell to signify whether the MD was in or out of the building. I remember the memory boards with transistors, capacitors, and resistors arranged neatly in rows. When the 3 digit number was entered a rather large stepping relay was cycled to access the correct memory cell. No logic chips here, just transistors and relays. What a great example of Dr. Dennard's memory cell operating outside of the computer environment!
As with my Commodore 64, I became intrigued with the innards of my PC-XT clone as I continued to look for new ways for improvement. I tired quickly of the constraints of my 360 kilobyte floppy drive and looked for ways to increase storage.. At this point two things were required. Since the main boards and BIOS were not designed to support a hard drive, I needed both a new hard drive and controller card. This would set me back about $300 to $400 which I defiantly could not afford. After searching around, I came across a great deal! A friend had a used 20 megabyte hard-card drive for sale for the price of $175.00. Hard-cards were one of the industries answers to a quick drive upgrade when the main board did not have hard drive support built in. The drive and controller were mounted together on a single 8 bit full length slot card. I can't remember what make of drive was installed, but I do remember the controller as being made by Western Digital. That coupled along with a new copy of MS-DOS 3.3 gave me plenty of drive space. There were larger hard drives around, but the 20 meg did nicely. It was IBM who built the first commercial hard drive back in 1956. It consist of fifty 24 inch platters and had a storage capacity of 5 megabytes. Hard drive and controller technology for older PC's was limited to MFM (Modified Frequency Modulation) and RLL (Run Length Limited) encoding. RLL was a modified version of MFM encoding that allowed for 50% more storage capacity on a drive. As hard drive size increased, many computer users were not ready open their wallets again for more drive space. A company called Perstor came up with a solution that was called ADRT (Advanced Data Recording Technology). With it they managed to increase the data storage on both MFM and RLL drives by 90% along with an increase in data transfer rate. A RLL drive could be used with a MFM controller, and if the MFM drive was a high quality drive, many times it would work well with an RLL controller increasing it's storage by 50%. But the Perstor ADRT controller would work with both the MFM and RLL drives. A standard MFM drive and controller usually had 17 sectors per track, a 5 megabit (Megabit not Megabyte) per second transfer rate and 32 bit error correction. An RLL drive and controller had 26 sectors per track, a 7.5 megabit data transfer rate, and 32 bit error correction. But, either a MFM or RLL drive mated with a Perstor ADRT controller would increase the sectors to 31 per track, increase the data transfer rate to 9 megabits per second, and used 56 bit error correction. I looked upon the Perstor board with the great interest, but again it was an issue of funds. Around six months later I bought a used 40 megabyte drive cheap that was having some problems. It turned out to be an easy fix, a poor solder connection on the power connector. The two drive 60 megabyte combination worked quite well for long time. Those early drives for PC's were monsters. Some of them were full height 5.24" drives and were as heavy as a couple of bricks! A far cry from the small light drives we have today.
It was not long before the hunt was on for more speed. The H/Z CPU board contained an Intel 8088-2 microprocessor that would run up to 8 MHZ clock speed, but of course not fast enough! NEC had developed an 8088 clone CPU which was called the V20. It was a direct replacement for the 8088-2 microprocessor and was designed to give a 20% boost in processing speed at the same clock speed. I soon placed an order from Jameco Electronics and soon I was pushing my old XT to it limit! Being curious, I checked to see if Jameco Electronics was still in business. Upon finding their web site I did a search for the V20. Believe it or not, they are still selling them, though they are used selling for $10.99.
With the expansion of the PC industry there were plenty of interesting pieces of software to experiment with. TSR (Terminate Stay Resident) programs were especially interesting. Before multitasking was commonplace for PC computers, there were many schemes and tricks to add different functions at the touch of a keystroke in our PC's. TSR's were programs that were loaded after DOS was started but before an application was started. The TSR started, installed itself in memory, communicated with DOS telling it not to over write the memory space, and then exited while hooking itself into the computers interrupt vectors waiting for a keystroke to reactivate itself. The computer constantly polls or cycles through these interrupt vectors, which upon certain predetermined conditions, turns control of the computer over to another piece of software, hardware, or the TSR. This is similar to us if we are reading the newspaper and the phone rings which interrupts us and we answer it. When done with the phone call we go back to reading the paper! With a quick keystroke combination, you could have the TSR become active and a pop up window would appear on the screen. My favorite TSR programs were those such as Borland's Sidekick, and two utility programs, Scout and PC Tools which was published by Central Point Software. Scout was a handy file management program that let you easily copy, delete, format disks etc. PC Tools was a full blown utility program that let you do just about anything file and disk management related plus recovery tools and a menu program. Similar to Norton Utilities, PC Tools was my favorite. Menu programs were similar to TSR's also. Before windows, you could us a menu program to manage and launch all your software. Upon launching a software title, the menu program would exit and leave a very small part of itself in memory. We always called this a footprint. The smaller the footprint the software left in memory the better. This would leave more memory available to run other software. After you exited the software title, the menu program would restore itself. My favorite was Quickmenu III because it sported a graphical desktop environment similar to windows and had built in disk & file tools.
All these cool TSR type programs came at a cost of memory usage. By the time you loaded in all your favorite TSR programs, you were left with limited memory to run other applications. How did we break the 640 kilobyte memory limit? Ah, enter EMS (Expanded Memory) cards. The EMS standard (LIM) was designed by Lotus, Intel, and Microsoft, a joint effort to provide enough memory to hold large amounts of data such as when using a spreadsheet program. Through a software driver, the expanded memory would be made available to programs through the trick of bank switching. The expanded memory board could hold up to a few megabytes of ram. The software driver would set up a 64 kilobyte window in the PC's memory, in 16 kilobyte increments called pages. Selections of EMS memory could be bank switched into the 64 kilobyte window in conventional memory and used by the program requesting it. This way a program had access to large amounts of ram to use on the fly! My expanded memory board contained 384 kilobytes of ram, but I later installed 2 megabytes of DRAM chips to fill it up. Now, if your TSR programs were written to use EMS memory, it would happily install itself in EMS memory freeing up conventional memory. In this configuration, you only lost 64 kilobytes of conventional memory plus the driver which had a small footprint.
By this time we would have replaced our monochrome screen with a CGA (Color Graphics Adaptor), EGA (Extended Graphics Adaptor), or VGA (Video Graphics Display) card and monitor. Finding a deal on a Commodore 1084 monitor (I always wanted an Amiga), I bought an EGA graphics card which worked nicely with the Commodore monitor. CGA could only display 4 colors with a screen resolution of 320x200. EGA increased that to 16 colors with a screen resolution of 640x350. VGA came along and displayed 265 colors with a screen resolution of 640x480. I'll bet everyone said the same thing after getting their first VGA adaptor and monitor, "how could it get any better than this!" Well it has, thanks to the advances in technology we can enjoy crisp clear high definition video on our PC's.
Modems had come a long way since those old 300 bps connections. My first experience introduction to online connectivity was with a Commodore 64 and a 300 bps Hayes compatible modem. In 1962, AT&T manufactured the Bell 103, the first commercial modem which connected at 300 bps (bits per second). Dennis Hayes invented the personal computer modem in 1977. His set of modem commands would be the standard for many years to come. Dial up connection speed has increased to 56.6 kbps, but it seems that with fast network connections, dial up modems may soon be obsolete. The speeds have jumped from 300, 1200, 2400, 4800, 9600, 28.8, 33.6, and finally 56.6kbps. How did we manage at a snails pace? Well for one thing, the program file size was much smaller. A large file on our Commodore 64 would take 30 to 60 minutes to transfer at 300bps. Most of the free files and shareware was distributed on local Computer Bulletin Boards. Many of us grew up using these local Bulletin Boards, or BBS's as we called them to exchange mail, program files, and discussion groups. At first they were only single line dial up system operated by a home computer buff, to large multi-line commercial systems. In a single line system, users were given a limited time limit each day so that others might have the opportunity to call in. I had a BBS running on my Commodore 64 with a 2400 bps modem, and had named it the Info BBS. A fellow by the name of Greg Puntz wrote a great BBS program for the C-64 called Color 64 BBS. It would push the C-64 to it's limit of 2400 bps, and could sense if the caller was using a C-64/128 or a PC and display the proper screen codes for each. At that time PC's had nice sized hard drives, but the C-64's hardware was not designed with a hard drive in mind. So, someone came up with a design to add a hard drive to the C-64. If you could afford it, for around $500.00 you would get a case the size of a PC-XT with a 20 mb hard drive and associated interface boards installed to interface to your C-64. It seemed like a big investment to me, but there were many die-hard C-64 users that set up BBS's using that huge external hard drive!
As dial up speed increased, so did the costs of those new modems. In the mean time, there were some custom solutions designed to solve the cost problem. One of my favorite PC BBS's was a multi-line BBS called Exec-pc. Exec-pc required a subscription fee to access its vast shareware and freeware library, and coupled with long distant phone fees connection costs were rather high. So, the search was on to find a low cost connection solution. Exex-pc had installed a number of modems on it's multiple line dial up system that would connect at 9600 bps using it's own proprietary hardware. Since standard 9600 bps modems were in the $250.00 range, I purchased one of these 9600 bps modems manufactured by CompuCom called a Speed Modem, for $90.00. The SpeedModem was a 9600 bps high-speed modem with CompuCom's proprietary CSP modulation protocol. At 9600 bps the file transfers seemed to fly! That was fine for Exec-Pc, but what about the many BBS's scattered across the country? How did we access them at a nominal cost rate? Who remembers Galaxy Starlink? It was a great idea that provided cheep access to BBS's all over the country. For a small subscription fee, Galaxy Starlink would set up access through your local Tymnet provider. Tymnet was a commercial computer network created by Tymshare Inc., and later was acquired by MCI. At the time, it was the largest commercial network in the United States with nodes in every major city. MCI has since terminated the Tymnet Network in 2002. Once your account was set up at Starlink, you could do one of two things. First you could log into Galaxy Starlink's main computer, which was a huge BBS type system. Second, you could log into your local Tymnet node, access a Tymnet node in a remote city, and actually dial out of the remote node to any local number toll free. This way you could access BBS's in any city that had a Tymnet node. Though the connection speed was limited to 1200/2400 bps, it was very cost effective. Galaxy Starlink charged by the amount of data that you actually moved through the system. At 1200 bps you could move approximately 200 kilobytes of data per hour when doing a file transfer for about $1.00. That might not seem like a lot of data, but remember the file sizes were smaller then, and dial up's like Quantum Link (which later became America On Line) were charging higher rates. I did not mind walking away from the computer for a few hours to download a number of files for only a few bucks! Even better, I could connect to a node in the city where Exec-pc was located and dial into it! Those were the days when it was really exciting to find ways to connect to a computer located half way across the country! How many people remember the early internet connections? I can't remember the name of my first internet provider, but I do remember it was text based, before any graphics were on the web pages! What excitement when first connecting to a web site in England! It was almost like you were there, reading data from a computer half way across the world! I understand how exciting it must have been when voice communication took place via the first transatlantic cable! Now it is so common place that we take our internet connection for granted. We can access information just about anywhere in the world via the internet.
Technology seems to have a way of being improvised to provide more speed, more features etc. Sometime after the 80286 CPU was released, someone came up with a cost effective idea geared towards upgrading a PC XT, the 286 XT main board. I had one of these peculiar main boards. It was a blend of XT technology using a 80286 CPU. It used a unique idea, a multi-bus system which gave it a big advantage over the standard XT board. The multi-bus system incorporate a fast 16 bit bus used for the CPU and RAM, and an 8 bit bus for expansion board I/O. The 16 bit bus , RAM, and onboard I/O was clocked at full 80286 speed, which was 12 Mhz on my board. The expansion bus was clocked at the standard 4.77 MHz XT speed. It could also access 1 Mb of memory. This was great because now using MSDOS 5, I could use the MSDOS EMM386 program and Quarterdeck's Qemm to load drivers and such in high memory, thus freeing up more memory for running programs. Remember the 64 kilobyte EMS page frame I mentioned before? Now, instead of having it located in low memory I could place it up in that mysterious high memory area freeing up 64 kilobyte of ram in low memory! For awhile this combination worked well for me, but as always, the need for speed gives reason to press on! Again, it was an issue of cost. How to upgrade your PC with minimal investment. The 80486 CPU had been on the market but the cost of a new 486 main board was out of reach for me. I was fortunate enough to find a 80386 main board that was being disposed of in favor of a 486. This was an earlier board that still had the old DRAM memory sockets. By that time, having acquired a number of older DRAM memory chips I was able to populate the RAM up to a whopping 8 megabytes! Wow, now I could use an extended memory manager (XMS) and access memory above the 1 megabyte limit that was imposed by my 286 based main board. Now it was time for some serious gaming! Up until then I had always used my C-64 for gaming, but things were changing again in the PC arena, and I now needed sound. In the early 80's Ken and Roberta Williams founded Sierra Online. They set the standard for graphical adventure games, which like myself, many others enjoyed on their C-64 computers. PC's in the late 80's had one drawback, no low cost quality gaming sound. In 1988 Sierra got together with a company called Adlib and used the Adlib music synthesizer standard in their games. My first sound card was an Adlib card, then later a Creative Labs Soundblaster.
By then I was experimenting with MS Windows. It was hard for all of us to give up MS DOS, it was like an old friend. I would guess that if you were find someone today who was proficient with MS DOS, and set them in front of an old PC, it would all come back to them as if they were picking up where they left off. It was not too long ago that one of the guys in the Biomedical department at the hospital where I work discovered that one of the old monitoring systems (which incidently never failed up until then) was running MS DOS! He came to me and said he need someone who knew what to do with DOS. Wow, it kinda felt good playing with an old DOS based PC! So armed with my copy of Spinrite, we resurrected the damaged hard drive and copied the software to a good hard drive, rebuilt the damaged boot up batch file and put it back in service! I don't remember what CPU was under the hood, but it must have been slow because it took Spinrite about 6 hours to work it's magic on the 200 Meg hard drive!
As always there were new toys to play with. Along came Windows 95 with it's hardware hungry requirements. When Windows 95 was released, it came on about a dozen 3.5" floppy disks. To fit more data on the floppy disks, Microsoft came up with a idea to add extra sectors to the disks. This could be accomplished by using high quality floppy disks and a special format. By doing this, Microsoft also thwarted attempts to copy the disks because the standard disk copy program would not re-produce the special format. It was not long before programmers released disk copy programs that not only could reproduce Microsoft's special format, but also let the user utilize the extra space for normal data storage. As I can recall, a high quality 3.5" 1.44 mb floppy could be formatted to around 1.6 mb of usable space.
So there I was again, trying to squeeze more from my hardware. Getting Windows 95 to run at a respectable speed on my old 386 main board was quite a feat. The early version of Windows 95 would run with only 8mb of system ram. It ran on my 386 main board, but very slow! Again my wallet presented a roadblock to upgrading. But, along came Evergreen with a great innovation called the 386 to 486 upgrade CPU. It plugged into a 386 socket, but contained an Intel 486 cpu which would clock up to 133 MHz! This enabled Windows 95 to run at a respectable speed on my old main board! Then the vicious hardware cycle started all over again! Faster video, better sound card, faster modem, and bigger hard drive to hold the data that Windows 95 needed. Then I got the granddaddy of all toys. Finally something that not too many of my frends had. One of the early CD ROM drives. My wife was doing a lot of horticulture slide photography at that time. She had an offer to have her slide collections put on CD, to be made available for sale. Well, as I explained it to her, we needed a CD-ROM to view the disks. I don't know how I talked her into that one, but that external Sony 1x CD-ROM reader set me back $500.00 ! Ah, but it was way cool like having a space shuttle in your living room! Not long after that the software companies started to release software, including games on CD which were great fun.
By then Microsoft had released Windows 95 on CD with many enhancements. Again, the need for speed required finding cost effective new hardware. Along came a company that challenged Intel with a cheaper fast CPU that fit my wallet nicely. NexGen developed an efficent RISC CPU and mainboard combo that would run at 90mhz and 120mhz, which was supposed to out perform the Intel pentium. It was a fast and cost effective way of acheving speed without the cost of a Pentium cpu. The cpu was not directly interchangeable with a Pentium, so NexGen sold it with their own mainboard. At that time, AMD had realized it had hit a brick wall with it's K5 cpu. AMD was in need an infusion of new technology to continue competing with Intel. They ended up purchasing the NexGen technology and used it in developing the AMD K6 cpu. So at that point, the technology seemed to advance pretty quickly. Moving from the K6 chips to the K6-II, k6-III, Windows 98, and finally into the Athlon's and Windows XP. As you can tell I have always been an AMD fan. I always had built my own systems until the past few years. Guess you could say "been there done that." I'm on my 2nd Dell Pentium based laptop, and am quite content with it. The first was a P4 2.6 ghz which I passed on to a family member, and right now I am typing this on an I6000 Centrino Pentium M widescreen laptop, nice!
01/03/2010 Update
Well, a lot has changed again since I wrote the above article. I have since given my Dell laptops away to family members, and have worked my way through a few more home brew computer configerations. I have moved away from Microsoft Windows, I just got tired of shelling out cash for their new operating systems. I have used a few favorers of Linux, and have settled with Ubuntu, just because it works. I have installed it on various PC's with out any major problems. So what am I using now? Well for portabliy, I have an Acer Aspire One Netbook, running Ubuntu. It works great. I just recently built a new box. I wanted a small low power PC, so I used an ASUS AT3N7A-I board that sports an Intel Atom processor with a Nvidia Ion chipset, and an Apex mini ITX case. I have read reviews of others putting down those Atom CPU's. My last mainboard had a 3ghz single core Intel CPU with an 800mhz bus, which at this point ran, fast with anything I threw at it. The Atom benchmarks faster than the 3ghz Pentium did. So I am happy, Ubuntu runs great on it.
machine created by man for his purposes. The existence of the personal computer seems to have become a part of our everyday life, intertwined with our day to day activities. Think about it, even if we choose not to own one, there is a computer involved somewhere, linking our choices to an outcome that has somehow been processed, analyzed, and crunched in a silicon brain. When the founder of digital Research, Dr. Gary Kildall wrote a primitive operation system called CP/M or Control Program for Microcomputers, it was designed to work only with IBM compatible soft sectored floppy disks. As the need arose to use CP/M with non IBM compatible floppy disks, Gary isolated the code that integrated the format of the disks to the rest of the operating system into a separate "BIOS" module that could be easily change to meet the requirements of other disk formats. The name stuck, the rest is history. This BIOS has evolved to be the life blood of our PC today. It seems to keep the hardwired components of our modern day PC working together in harmony. Does this scenario sound familiar? Man always strives mimic what God has created.
In 1987-88, while oblivious to any other ground breaking developments in technology, I was happily delving into the innards and secrets of my Commodore 64. After all, the secrets it held were captivating to no end! Before that I tinkered with an APF M1000. It sported a Motorola 6800 microprocessor, built in basic interpreter, and 8 kilobytes of system RAM (Random Access Memory) which I soon modified to 24 kilobytes! It wasn't until I took a home study course in computer technology, that I was introduced to the IBM clones. As part of the course, the lucky student built a brand new shinny Heath/Zenith clone PC complete with DOS 2.0! After it was all said and done, what was I to do with a computer with a monochrome screen, and only 128 kilobytes of memory? At that time memory was expensive (to me anyway). So shelling out the cash to bring the memory up to a whopping 512 kilobytes gave me no satisfaction. Then I found a whole new world of PC technology beyond the excitement that my Commodore 64 could provide to captivate my attention, and have been hooked ever since.
Early on, there were three main driving forces that shaped the IBM compatible PC industry, IBM, Intel and Microsoft. Of course there were plenty others before and after, who contributed immensely, and they are still contributing to the industry today. But, together IBM, Intel and Microsoft linked the ground working tools that the industry first embraced on a large scale to get the ball rolling. Microsoft DOS was actually a clone of CP/M. When IBM's interest in using the CP/M operating system in their first personal computer was put off by Gary Kildall, they turned to Microsoft for an operating system for the computer that would change the way we live. William Gates at Microsoft saw an opportunity waiting, and agreed to provide an operating system for IBM's new child. Microsoft having no solution of it's own, turned to Seattle Computer Products. Early in 1980, Tim Paterson, who worked for Seattle Computer products, had written a cloned version of CP/M called 86-DOS to run on Seattle's S-100 bus 8086 based microcomputer. Up until that time, CP/M had not been ported to the Intel 8086 microprocessor platform. So Microsoft licensed 86-DOS from Seattle Computer and presented it to IBM as Microsoft Disk Operating System 1.0. Since IBM could not find any other suitable offerings, they agreed to strike a deal with Microsoft to use MS-DOS in their new 8086 based PC. So in July 1981, Microsoft purchased all rights to 86-DOS, and in August 1981, IBM announced the IBM PC Model 5150, which became available for purchase in the fall of 1981. The PC business was a good run for IBM, 23 years. In the fall of 2004 IBM announced it would sell it's PC division to a Chinese manufacture, Lenovo.
If you were to ask who is responsible for this invention that has change our lives, I could only answer 'no one in particular.' It were those with vision that built upon, and combined the work of others to create what we have today. How far back shall we look for the roots of the PC? It is what we experience that we remember the most. We program software to run on a specific computer platform. But we are programmed by experience. It seems that we were created to learn from our past mistakes and accomplishments, and desires. Shall we remember Dr. Ted Hoff, who in 1971 created the Intel 4004, which was the first commercially available microprocessor. I remember the wonder in reading about the Altair 8800 computer in Popular Electronics. If you had enough money to spare, you could actually buy it in kit form. Two college students, William Gates, and Paul Allen wrote the BASIC language interpreter for the Altair 8800, and later formed Microsoft to sell their software product. And let's not forget about Steve Wozniak and Steve Jobs who built the Apple computer.
It wasn't until I pursued my first home study computer course in 1980 that I entered into the world of digital logic circuits. Similar to the Altair 8800, the finished product of the course that I took was a good sized box with many switches and lights with which my new toy was programmed. Computers then were in their most basic form, 0's and 1's, digital logic. How many of us every day ignore the fact that turning on the light switch in our homes is a digital function, off is '0', on is '1'. It is the same as our PC's today do it, except they manipulate the 0's and 1's elegantly on a far greater scale. In high school I created a electrical tic-tac-toe game. It was driven by relays, dozens of them. It was hardwired logic, and un-beatable as long as it was allowed to make the first move. I was only one person whose creation only served myself. But what about the countless others that have made small contributions similar to my logic driven invention, which were used to build upon, to create what we have today. It is said that ENIAC, the first vacuum tube computer was contained in a room 30x50 feet, weighed 30 tons, and contained 17,468 vacuum tubes. It could executed 5,000 additions and 300 multiplications per second. In contrast the Intel 4004, an early microprocessor contained 2300 transistors and executed 60,000 instructions per second.
How fast shall we go? Steve Gibson, founder of Gibson Research, wrote a column for Infoworld magazine for a number of years. In 1989, he wrote something that no one including himself, ever would have thought could happen. He wrote, " We've grown so accustomed to the clock rate ride that it's easy to simply anticipate that the joy-ride will continue through coming years with announcements of 40, 45, 50, 65, 75, 100, 150 megahertz systems. It won't happen. You're sons will not be bouncing their grandsons upon their knees telling them of the days when PC's didn't all run at 12 gigahertz. Just as the performance of dynamic RAM memory bottomed out several years ago, we've come to the end of our years of clock rate joy riding... completely regardless of how much we're willing to pay. " I imagine he is chuckling about that statement today. Even though the industry has found ways to speed up microprocessors at lower clock speeds, we are well into the 4 gigahertz range. How long until we hit that 12 gigahertz mark?
I started my venture into the PC word with a Heath/Zenith PC-XT 8088 8 megahertz clone when the Intel 80386 CPU was arriving on the scene. Heath/Zenith took a different approach to design compared to the standard today. Instead of a main board, they designed a backplane board which all the other boards plugged into. As I recall, the CPU, and memory sat on a single board, The I/O (Input Output) functions on a separate board, and also the video on a separate board. The 360 kilobyte 5 1/4" floppy drive held more than that of my Commodore 64, but I had and no sound, a green monochrome screen, and no hard drive. How did I ever survive that? Computer parts were quite expensive at that time, so I was always a few steps behind on the technology. This gave me the motivation to learn how to squeeze every last ounce of speed out of my hardware. If I were to teach a class or give a lecture on using computers, I would most likely find it difficult, unless I were permitted to explain the bowels of these perplexing machines. It seems that today some of the fun has diminished because now we are able to quickly boot up our PC's, and open an abundance of graphical windows while also doing a multitude of other things (Multitasking), most of which is not prominent or readily noticeable to us. We have plenty of RAM, plenty of hard disk space, plenty of resources. This is good news for the general population of computer users because it relinquishes more time to focus on the task at hand. But also bad for inquisitive hardware creatures like myself just because things are running just to darn smooth lately! In those days (not that long ago) I envied anyone who could afford to keep up with the technology with all the latest computer gadgets. In my circle of computer enthusiasts the driving force was how can we get more while spending less! I soon discovered that unlike my trusty Commodore 64, this inscrutable PC seemed to have unlimited potential to expand. So there I was, looking at empty sockets, how tempting it was to easily increase the RAM! After assembly, my PC-XT clone which only sported 128 kilobytes of RAM. To increase it in any amount required a set of nine identical DRAM (Dynamic Random Access Memory) chips. Unlike today's memory module, each chip was flat with downward pins on each side which fit nicely into a socket. Having populated my sockets with nine more 128 kilobyte, and nine more 256 kilobyte chips, I now could enjoy 512 kilobytes of system RAM. I suspect Dr. Robert H. Dennard, the inventor of DRAM may have not realized the extent of his creation. Working for IBM at the time, he had an idea to design a simple memory cell using a single transistor and capacitor, a device similar to a short term storage battery. His memory cells, could be accessed or written to independently of each other. Remember those 0's and 1's? Each cell in a bank of memory cells can be changed to '0' - off, or '1' - on. For many years I have worked in the Engineering Department of a local hospital. In doing so I have been given the opportunity to see first hand much of computer history, including Dr. Dennard's invention in action. Presently we have many ways to keep in touch and keep track of each other. But in the past, the telephone switchboard consisted of rows of jacks and plugs to manually connect calls to their destination. In those days, thanks to Dr. Dennard's invention, when entering the building our MD's entered a 3 digit code into a key pad while pressing either an IN or OUT button. This simply programed one of Dr. Dennard's memory cells either on or off. Accordingly an indicator light referenced the state of the memory cell to signify whether the MD was in or out of the building. I remember the memory boards with transistors, capacitors, and resistors arranged neatly in rows. When the 3 digit number was entered a rather large stepping relay was cycled to access the correct memory cell. No logic chips here, just transistors and relays. What a great example of Dr. Dennard's memory cell operating outside of the computer environment!
As with my Commodore 64, I became intrigued with the innards of my PC-XT clone as I continued to look for new ways for improvement. I tired quickly of the constraints of my 360 kilobyte floppy drive and looked for ways to increase storage.. At this point two things were required. Since the main boards and BIOS were not designed to support a hard drive, I needed both a new hard drive and controller card. This would set me back about $300 to $400 which I defiantly could not afford. After searching around, I came across a great deal! A friend had a used 20 megabyte hard-card drive for sale for the price of $175.00. Hard-cards were one of the industries answers to a quick drive upgrade when the main board did not have hard drive support built in. The drive and controller were mounted together on a single 8 bit full length slot card. I can't remember what make of drive was installed, but I do remember the controller as being made by Western Digital. That coupled along with a new copy of MS-DOS 3.3 gave me plenty of drive space. There were larger hard drives around, but the 20 meg did nicely. It was IBM who built the first commercial hard drive back in 1956. It consist of fifty 24 inch platters and had a storage capacity of 5 megabytes. Hard drive and controller technology for older PC's was limited to MFM (Modified Frequency Modulation) and RLL (Run Length Limited) encoding. RLL was a modified version of MFM encoding that allowed for 50% more storage capacity on a drive. As hard drive size increased, many computer users were not ready open their wallets again for more drive space. A company called Perstor came up with a solution that was called ADRT (Advanced Data Recording Technology). With it they managed to increase the data storage on both MFM and RLL drives by 90% along with an increase in data transfer rate. A RLL drive could be used with a MFM controller, and if the MFM drive was a high quality drive, many times it would work well with an RLL controller increasing it's storage by 50%. But the Perstor ADRT controller would work with both the MFM and RLL drives. A standard MFM drive and controller usually had 17 sectors per track, a 5 megabit (Megabit not Megabyte) per second transfer rate and 32 bit error correction. An RLL drive and controller had 26 sectors per track, a 7.5 megabit data transfer rate, and 32 bit error correction. But, either a MFM or RLL drive mated with a Perstor ADRT controller would increase the sectors to 31 per track, increase the data transfer rate to 9 megabits per second, and used 56 bit error correction. I looked upon the Perstor board with the great interest, but again it was an issue of funds. Around six months later I bought a used 40 megabyte drive cheap that was having some problems. It turned out to be an easy fix, a poor solder connection on the power connector. The two drive 60 megabyte combination worked quite well for long time. Those early drives for PC's were monsters. Some of them were full height 5.24" drives and were as heavy as a couple of bricks! A far cry from the small light drives we have today.
It was not long before the hunt was on for more speed. The H/Z CPU board contained an Intel 8088-2 microprocessor that would run up to 8 MHZ clock speed, but of course not fast enough! NEC had developed an 8088 clone CPU which was called the V20. It was a direct replacement for the 8088-2 microprocessor and was designed to give a 20% boost in processing speed at the same clock speed. I soon placed an order from Jameco Electronics and soon I was pushing my old XT to it limit! Being curious, I checked to see if Jameco Electronics was still in business. Upon finding their web site I did a search for the V20. Believe it or not, they are still selling them, though they are used selling for $10.99.
With the expansion of the PC industry there were plenty of interesting pieces of software to experiment with. TSR (Terminate Stay Resident) programs were especially interesting. Before multitasking was commonplace for PC computers, there were many schemes and tricks to add different functions at the touch of a keystroke in our PC's. TSR's were programs that were loaded after DOS was started but before an application was started. The TSR started, installed itself in memory, communicated with DOS telling it not to over write the memory space, and then exited while hooking itself into the computers interrupt vectors waiting for a keystroke to reactivate itself. The computer constantly polls or cycles through these interrupt vectors, which upon certain predetermined conditions, turns control of the computer over to another piece of software, hardware, or the TSR. This is similar to us if we are reading the newspaper and the phone rings which interrupts us and we answer it. When done with the phone call we go back to reading the paper! With a quick keystroke combination, you could have the TSR become active and a pop up window would appear on the screen. My favorite TSR programs were those such as Borland's Sidekick, and two utility programs, Scout and PC Tools which was published by Central Point Software. Scout was a handy file management program that let you easily copy, delete, format disks etc. PC Tools was a full blown utility program that let you do just about anything file and disk management related plus recovery tools and a menu program. Similar to Norton Utilities, PC Tools was my favorite. Menu programs were similar to TSR's also. Before windows, you could us a menu program to manage and launch all your software. Upon launching a software title, the menu program would exit and leave a very small part of itself in memory. We always called this a footprint. The smaller the footprint the software left in memory the better. This would leave more memory available to run other software. After you exited the software title, the menu program would restore itself. My favorite was Quickmenu III because it sported a graphical desktop environment similar to windows and had built in disk & file tools.
All these cool TSR type programs came at a cost of memory usage. By the time you loaded in all your favorite TSR programs, you were left with limited memory to run other applications. How did we break the 640 kilobyte memory limit? Ah, enter EMS (Expanded Memory) cards. The EMS standard (LIM) was designed by Lotus, Intel, and Microsoft, a joint effort to provide enough memory to hold large amounts of data such as when using a spreadsheet program. Through a software driver, the expanded memory would be made available to programs through the trick of bank switching. The expanded memory board could hold up to a few megabytes of ram. The software driver would set up a 64 kilobyte window in the PC's memory, in 16 kilobyte increments called pages. Selections of EMS memory could be bank switched into the 64 kilobyte window in conventional memory and used by the program requesting it. This way a program had access to large amounts of ram to use on the fly! My expanded memory board contained 384 kilobytes of ram, but I later installed 2 megabytes of DRAM chips to fill it up. Now, if your TSR programs were written to use EMS memory, it would happily install itself in EMS memory freeing up conventional memory. In this configuration, you only lost 64 kilobytes of conventional memory plus the driver which had a small footprint.
By this time we would have replaced our monochrome screen with a CGA (Color Graphics Adaptor), EGA (Extended Graphics Adaptor), or VGA (Video Graphics Display) card and monitor. Finding a deal on a Commodore 1084 monitor (I always wanted an Amiga), I bought an EGA graphics card which worked nicely with the Commodore monitor. CGA could only display 4 colors with a screen resolution of 320x200. EGA increased that to 16 colors with a screen resolution of 640x350. VGA came along and displayed 265 colors with a screen resolution of 640x480. I'll bet everyone said the same thing after getting their first VGA adaptor and monitor, "how could it get any better than this!" Well it has, thanks to the advances in technology we can enjoy crisp clear high definition video on our PC's.
Modems had come a long way since those old 300 bps connections. My first experience introduction to online connectivity was with a Commodore 64 and a 300 bps Hayes compatible modem. In 1962, AT&T manufactured the Bell 103, the first commercial modem which connected at 300 bps (bits per second). Dennis Hayes invented the personal computer modem in 1977. His set of modem commands would be the standard for many years to come. Dial up connection speed has increased to 56.6 kbps, but it seems that with fast network connections, dial up modems may soon be obsolete. The speeds have jumped from 300, 1200, 2400, 4800, 9600, 28.8, 33.6, and finally 56.6kbps. How did we manage at a snails pace? Well for one thing, the program file size was much smaller. A large file on our Commodore 64 would take 30 to 60 minutes to transfer at 300bps. Most of the free files and shareware was distributed on local Computer Bulletin Boards. Many of us grew up using these local Bulletin Boards, or BBS's as we called them to exchange mail, program files, and discussion groups. At first they were only single line dial up system operated by a home computer buff, to large multi-line commercial systems. In a single line system, users were given a limited time limit each day so that others might have the opportunity to call in. I had a BBS running on my Commodore 64 with a 2400 bps modem, and had named it the Info BBS. A fellow by the name of Greg Puntz wrote a great BBS program for the C-64 called Color 64 BBS. It would push the C-64 to it's limit of 2400 bps, and could sense if the caller was using a C-64/128 or a PC and display the proper screen codes for each. At that time PC's had nice sized hard drives, but the C-64's hardware was not designed with a hard drive in mind. So, someone came up with a design to add a hard drive to the C-64. If you could afford it, for around $500.00 you would get a case the size of a PC-XT with a 20 mb hard drive and associated interface boards installed to interface to your C-64. It seemed like a big investment to me, but there were many die-hard C-64 users that set up BBS's using that huge external hard drive!
As dial up speed increased, so did the costs of those new modems. In the mean time, there were some custom solutions designed to solve the cost problem. One of my favorite PC BBS's was a multi-line BBS called Exec-pc. Exec-pc required a subscription fee to access its vast shareware and freeware library, and coupled with long distant phone fees connection costs were rather high. So, the search was on to find a low cost connection solution. Exex-pc had installed a number of modems on it's multiple line dial up system that would connect at 9600 bps using it's own proprietary hardware. Since standard 9600 bps modems were in the $250.00 range, I purchased one of these 9600 bps modems manufactured by CompuCom called a Speed Modem, for $90.00. The SpeedModem was a 9600 bps high-speed modem with CompuCom's proprietary CSP modulation protocol. At 9600 bps the file transfers seemed to fly! That was fine for Exec-Pc, but what about the many BBS's scattered across the country? How did we access them at a nominal cost rate? Who remembers Galaxy Starlink? It was a great idea that provided cheep access to BBS's all over the country. For a small subscription fee, Galaxy Starlink would set up access through your local Tymnet provider. Tymnet was a commercial computer network created by Tymshare Inc., and later was acquired by MCI. At the time, it was the largest commercial network in the United States with nodes in every major city. MCI has since terminated the Tymnet Network in 2002. Once your account was set up at Starlink, you could do one of two things. First you could log into Galaxy Starlink's main computer, which was a huge BBS type system. Second, you could log into your local Tymnet node, access a Tymnet node in a remote city, and actually dial out of the remote node to any local number toll free. This way you could access BBS's in any city that had a Tymnet node. Though the connection speed was limited to 1200/2400 bps, it was very cost effective. Galaxy Starlink charged by the amount of data that you actually moved through the system. At 1200 bps you could move approximately 200 kilobytes of data per hour when doing a file transfer for about $1.00. That might not seem like a lot of data, but remember the file sizes were smaller then, and dial up's like Quantum Link (which later became America On Line) were charging higher rates. I did not mind walking away from the computer for a few hours to download a number of files for only a few bucks! Even better, I could connect to a node in the city where Exec-pc was located and dial into it! Those were the days when it was really exciting to find ways to connect to a computer located half way across the country! How many people remember the early internet connections? I can't remember the name of my first internet provider, but I do remember it was text based, before any graphics were on the web pages! What excitement when first connecting to a web site in England! It was almost like you were there, reading data from a computer half way across the world! I understand how exciting it must have been when voice communication took place via the first transatlantic cable! Now it is so common place that we take our internet connection for granted. We can access information just about anywhere in the world via the internet.
Technology seems to have a way of being improvised to provide more speed, more features etc. Sometime after the 80286 CPU was released, someone came up with a cost effective idea geared towards upgrading a PC XT, the 286 XT main board. I had one of these peculiar main boards. It was a blend of XT technology using a 80286 CPU. It used a unique idea, a multi-bus system which gave it a big advantage over the standard XT board. The multi-bus system incorporate a fast 16 bit bus used for the CPU and RAM, and an 8 bit bus for expansion board I/O. The 16 bit bus , RAM, and onboard I/O was clocked at full 80286 speed, which was 12 Mhz on my board. The expansion bus was clocked at the standard 4.77 MHz XT speed. It could also access 1 Mb of memory. This was great because now using MSDOS 5, I could use the MSDOS EMM386 program and Quarterdeck's Qemm to load drivers and such in high memory, thus freeing up more memory for running programs. Remember the 64 kilobyte EMS page frame I mentioned before? Now, instead of having it located in low memory I could place it up in that mysterious high memory area freeing up 64 kilobyte of ram in low memory! For awhile this combination worked well for me, but as always, the need for speed gives reason to press on! Again, it was an issue of cost. How to upgrade your PC with minimal investment. The 80486 CPU had been on the market but the cost of a new 486 main board was out of reach for me. I was fortunate enough to find a 80386 main board that was being disposed of in favor of a 486. This was an earlier board that still had the old DRAM memory sockets. By that time, having acquired a number of older DRAM memory chips I was able to populate the RAM up to a whopping 8 megabytes! Wow, now I could use an extended memory manager (XMS) and access memory above the 1 megabyte limit that was imposed by my 286 based main board. Now it was time for some serious gaming! Up until then I had always used my C-64 for gaming, but things were changing again in the PC arena, and I now needed sound. In the early 80's Ken and Roberta Williams founded Sierra Online. They set the standard for graphical adventure games, which like myself, many others enjoyed on their C-64 computers. PC's in the late 80's had one drawback, no low cost quality gaming sound. In 1988 Sierra got together with a company called Adlib and used the Adlib music synthesizer standard in their games. My first sound card was an Adlib card, then later a Creative Labs Soundblaster.
By then I was experimenting with MS Windows. It was hard for all of us to give up MS DOS, it was like an old friend. I would guess that if you were find someone today who was proficient with MS DOS, and set them in front of an old PC, it would all come back to them as if they were picking up where they left off. It was not too long ago that one of the guys in the Biomedical department at the hospital where I work discovered that one of the old monitoring systems (which incidently never failed up until then) was running MS DOS! He came to me and said he need someone who knew what to do with DOS. Wow, it kinda felt good playing with an old DOS based PC! So armed with my copy of Spinrite, we resurrected the damaged hard drive and copied the software to a good hard drive, rebuilt the damaged boot up batch file and put it back in service! I don't remember what CPU was under the hood, but it must have been slow because it took Spinrite about 6 hours to work it's magic on the 200 Meg hard drive!
As always there were new toys to play with. Along came Windows 95 with it's hardware hungry requirements. When Windows 95 was released, it came on about a dozen 3.5" floppy disks. To fit more data on the floppy disks, Microsoft came up with a idea to add extra sectors to the disks. This could be accomplished by using high quality floppy disks and a special format. By doing this, Microsoft also thwarted attempts to copy the disks because the standard disk copy program would not re-produce the special format. It was not long before programmers released disk copy programs that not only could reproduce Microsoft's special format, but also let the user utilize the extra space for normal data storage. As I can recall, a high quality 3.5" 1.44 mb floppy could be formatted to around 1.6 mb of usable space.
So there I was again, trying to squeeze more from my hardware. Getting Windows 95 to run at a respectable speed on my old 386 main board was quite a feat. The early version of Windows 95 would run with only 8mb of system ram. It ran on my 386 main board, but very slow! Again my wallet presented a roadblock to upgrading. But, along came Evergreen with a great innovation called the 386 to 486 upgrade CPU. It plugged into a 386 socket, but contained an Intel 486 cpu which would clock up to 133 MHz! This enabled Windows 95 to run at a respectable speed on my old main board! Then the vicious hardware cycle started all over again! Faster video, better sound card, faster modem, and bigger hard drive to hold the data that Windows 95 needed. Then I got the granddaddy of all toys. Finally something that not too many of my frends had. One of the early CD ROM drives. My wife was doing a lot of horticulture slide photography at that time. She had an offer to have her slide collections put on CD, to be made available for sale. Well, as I explained it to her, we needed a CD-ROM to view the disks. I don't know how I talked her into that one, but that external Sony 1x CD-ROM reader set me back $500.00 ! Ah, but it was way cool like having a space shuttle in your living room! Not long after that the software companies started to release software, including games on CD which were great fun.
By then Microsoft had released Windows 95 on CD with many enhancements. Again, the need for speed required finding cost effective new hardware. Along came a company that challenged Intel with a cheaper fast CPU that fit my wallet nicely. NexGen developed an efficent RISC CPU and mainboard combo that would run at 90mhz and 120mhz, which was supposed to out perform the Intel pentium. It was a fast and cost effective way of acheving speed without the cost of a Pentium cpu. The cpu was not directly interchangeable with a Pentium, so NexGen sold it with their own mainboard. At that time, AMD had realized it had hit a brick wall with it's K5 cpu. AMD was in need an infusion of new technology to continue competing with Intel. They ended up purchasing the NexGen technology and used it in developing the AMD K6 cpu. So at that point, the technology seemed to advance pretty quickly. Moving from the K6 chips to the K6-II, k6-III, Windows 98, and finally into the Athlon's and Windows XP. As you can tell I have always been an AMD fan. I always had built my own systems until the past few years. Guess you could say "been there done that." I'm on my 2nd Dell Pentium based laptop, and am quite content with it. The first was a P4 2.6 ghz which I passed on to a family member, and right now I am typing this on an I6000 Centrino Pentium M widescreen laptop, nice!
01/03/2010 Update
Well, a lot has changed again since I wrote the above article. I have since given my Dell laptops away to family members, and have worked my way through a few more home brew computer configerations. I have moved away from Microsoft Windows, I just got tired of shelling out cash for their new operating systems. I have used a few favorers of Linux, and have settled with Ubuntu, just because it works. I have installed it on various PC's with out any major problems. So what am I using now? Well for portabliy, I have an Acer Aspire One Netbook, running Ubuntu. It works great. I just recently built a new box. I wanted a small low power PC, so I used an ASUS AT3N7A-I board that sports an Intel Atom processor with a Nvidia Ion chipset, and an Apex mini ITX case. I have read reviews of others putting down those Atom CPU's. My last mainboard had a 3ghz single core Intel CPU with an 800mhz bus, which at this point ran, fast with anything I threw at it. The Atom benchmarks faster than the 3ghz Pentium did. So I am happy, Ubuntu runs great on it.
