It catches your eye: Big-Top carnival writing on a 12-foot banner stretched between two wooden poles. In the city this would be interesting--but in the middle of a barren field, it's just plain odd. The banner reads:

Welcome to The Oldskool PC Carnival Sideshow!

Come witness electronic freaks of nature the likes of modern man has never seen!

Witness bizarre methods of PC software distribution -- ROM cartridges! UPC Barcodes!
Listen to the strange sounds of software add-on audio cassettes!
Stare at the twisted results of grotesque PC hardware experiments -- Rubber keyboards! Built-in monitors! Built-in printers!
Gawk abashedly at hardware-assisted compression for hard drives!

Come one, Come all!

In the distance behind the banner, you see a tall man standing on a soap box wearing a loud purple tuxedo, top hat, and cane. His eyes flash bright yellow as the sunlight reflects off his glasses; his coattails flap in the breeze. Even more strange is the tent behind him, with verticle yellow-and-white stripes, and a single wire running from a nearby power line to the topmost point of the tent.

"This is too weird," you think to yourself. "Someone must be filming a movie."

You park the car on the side of the road and walk up to the man. He watches you the entire way, then smiles excitedly, a twinkle behind his eyes. He steps down from his soap box and motions for you to come closer. "I'm glad you stopped," he says quickly. "This is probably your only chance to meet these lost children from the dawn of the computer age." He motions you closer, and you both walk to the entrance of the tent. "Sure, some people find them grotesque, but we owe it to ourselves to learn of their existence. What would happen otherwise? Would they cease to exist if nobody knew about them? Who would tell their stories?"

You both stop at the open flap of the tent. It's dark in there, and while you cannot directly see what's inside, you can hear the whisper and hum of electric fans and transformers, and can smell the faint odor of ozone. He notices your concentration and says enthusiastically, "Isn't it a twisted little symphony? Go on in! Immerse yourself!"

You walk in.

Welcome to some of the more interesting experiments--and dismal failures--of the early PC industry.

At the beginning of the personal computing industry, there were many attempts at creating the better mouse-trap. There was no immediate domineering computing standard, so general havoc ensued while companies strived to create the next big breakthrough product to become a standard--what would become the next Hayes modem, or perhaps the next Microsoft mouse, as we now know them.

Looking at PCs today, you can see established standards that we rely on, like hard drives, floppy disks, mice, network cards, and keyboards. They've been around as long as you can remember. But did you know that these peripherals were not always the standard on the PC? IBM had no guarantee that their heralded PC would become the dominating home platform, so even though it was doing well, IBM (and other PC-affiliated companies) experimented with new designs and ideas, just to be safe. A few of these ideas panned out; most did not. For example, equipping PCs with a 3.5" disk drive was an experiment on IBM's part that succeeded, and brought 3.5" disks for PCs into the mainstream. Their cartridge system incorporated into the PCjr, however, was misguided from the start.

My task in this essay is to bring those failures to light so that we can not only learn from them, but have a fun time doing so. While this is mostly entertaining, be prepared for a few other emotions surfacing along the way... sadness, for the good ideas that somehow failed, pity for the ideas that never had a chance, and sometimes utter confusion for the "What were they thinking?" products.

Grab a sandwich. Let's begin.

ROM Cartridges

because

"They're faster and more durable than floppy disks."

History

Cartridges, for those too young to remember, were very similar to video game cartridges (just like the Nintendo 64 game carts of today). They were usually made up of a single ROM chip encased in a hard plastic shell. Metal "edge" connectors were on one end to interface to the slot you inserted it into. Simple design--very durable, and they loaded programs quicker than floppy disks or tape because they literally became part of the computer's ROM once inserted.

Rewind to 1983: IBM decides that a home computer based on the widely-successful PC would be profitable. Looking around the other successful low-cost home computers at the time, most had a cartridge slot. I can just imagine the development meeting: "Atari, Timex, Texas Instruments, Commodore--they've all got cartridge slots, so we need a cartridge slot. Hell, the Atari 800 has two of them! Ours should as well!"

The PCjr, when finally introduced in 1984, had exactly that: Two cartridge slots, side by side, underneath the floppy drive bay. Cartridges contained a single 64K ROM, and could be inserted into either slot. Once inserted, the computer would reboot, and if necessary, "boot" the cart just like booting a floppy disk.

Why two slots? 64K was not enough space to hold then-complex PC programs, so it was assumed that any large program would need to spill over onto two carts.

PC Software Carts

The following is, to my best knowledge, a complete listing of all the software that was released for the PCjr's cartridge slot(s). It is based off of my personal memory, an IBM catalog from 1985, and some listings found on the web.

Verdict

A bad idea. Floppy disks held more data and were much cheaper, even if floppy drives were more expensive.

More Information

The Oldskool PC Shrine to the Tandy and PCjr expands on this and other PCjr oddities.

UPC Barcodes

because

"Magazines are cheaper than disks."

History

Some background: Whenever you go to the checkout counter of your local grocery store, your purchases are rung up by a scanner. The checkout clerk just runs your purchases over the scanner, and they're magically tallied in the cash register. This works because the scanner picks up the item number from a barcode on the product. You know--that little pattern of black and white vertical lines? More than just a pretty design, they are an encoding of numbers and letters in a form that the scanner can read with a low-power laser. It reads the thick and thin bars, translates the patterns back into numbers and letters, and then feeds them into a computer where your purchase is recognized and rung up correctly by the cash register.

The point of all this is that barcodes offer a quick, easy way to input data into a computer. Did a light go on in your head? It certainly did for Robert L. Brass, who founded Cauzin Systems, Inc. in 1985. Cauzin Systems was created to build and market the Cauzin SoftStrip System for personal computers, which debuted in October of that year. Along with Richard K. Balaska, Jr., Cauzin's primary engineer, The Cauzin SoftStrip System was a long barcode scanner that connected to a computer's serial port (or, in the Apple II's case, the cassette ports). The software that came with it could drive the scanner to read long barcodes, translate them back into the numbers and letters they represented, and save the whole thing to a file. Another program, released later, let you create your own barcodes by converting files into barcodes and printing them out on your dot matrix or laser printer.

Each barcode "strip" was about 8 inches long by about one inch wide. About 1K of information fit into each strip as printed on a 24-pin dot matrix printer, but a laser printer could cram about 2K into each strip. If you sent the material to SoftStrip, they could fit about 5K into a single strip due to their higher-grade printing process. About four strips could fit onto a single 8.5" by 11" piece of paper. Because many programs needed several barcode strips to be properly represented, Cauzin coined the term "Stripware" to describe a full page of strips. ("Strip software"--get it?) Of course, these are all informal measurements and specifications from my rotting memory; the SoftStrip Laser StripMaker manual says that for a 300 DPI desktop laser printer, the densities are approximately as follows:

• Low density: 590 bytes per 100mm
• Medium density: 1040 bytes per 100mm
• High density: 1630 bytes per 100mm

Note: Maximum strip length for vertical strips is 240mm and for horizontal strips 175mm.

Why would anyone want to do this? The basic idea behind the system was that it gave people an easy way to load the many BASIC program listings that were popular in magazines at the time without retyping them--simply print a barcode strip here and there in the margins of the magazine, and users are relieved from typing in hundreds of lines of code. The appeal to magazine publishers was that it added value to the magazine without the additional cost of distributing a floppy disk along with it. There were other minor benefits; for example, "Stripware" could survive magnetic disasters because it wasn't stored on magnetic media, but just regular paper.

The Cauzin SoftStrip System was interesting enough to win a MacUser magazine "Eddy" award for "Most Innovative Concept" in 1986. But by that point, it was failing miserably in the marketplace. By 1987, it was off the market, and no magazines were publishing listings in SoftStrip format.

With so many good things going for it, why did it fail? While some magazines like Family Computing, II computing, InCider, and even the prominent Byte Magazine gave it a shot, most magazine publishers were unwilling to dedicate space to the barcode strips (it sometimes took more than one page of strips to represent a large program, since the strips were relatively low density to account for the poor publishing/printing processes of the time). Nibble Magazine wouldn't touch it at all because they sold diskettes with the magazine's code listings separately for a fee; printing SoftStrip barcodes would have eliminated those sales. There was one other problem: Price. The SoftStrip Reader had a street price of about $200, which was beyond the casual purchase price of most computer consumers. Other companies tried their versions of barcode readers ("OSCAR", etc.), but at that time, any bar code reader that was cheap enough for home use would not read the bar codes reliably, so these devices failed as well.

Roy M. Silvernail has a more sinister theory as to why the SoftStrip System failed:

"Cauzin was bought by Eastman Kodak shortly after they started advertising in the big rags (like Byte, which printed a test strip a month or so before Cauzin made their big advertising push). Cauzin was even putting freeware programs in their ads. At that time, Kodak was just entering the soon-to-be-lucrative magnetic media market with their floppy disk line. I'd say they were bought and buried."

Verdict

A bad idea. Cauzin lacked the foresight to realize that a clever idea is not always practical and affordable.

More Information

Cauzin on The Computer Chronicles

Bill Gates mentioned that one of his earlier projects attempted to do the same thing--use a barcode scanner for personal computer use--in part of a Smithsonian Museum interview.

James Willing, in his excellent Computer Garage Museum, has pictures and a description of his SoftStrip Reader.

(in)Famous game programmer John Romero told CNET that one of his early games, Bongo's Bash, was created "and given away as a promotional item for... the little-known Cauzin Softstrip". It appeared in the Cauzin advertisements.

If you like, you can pick up a copy of the book "ANIMATED ALGORITHMS: A SELF-TEACHING COURSE IN DATA STRUCTURES AND FUNDAMENTAL ALGORITHMS" by Michael J. and Simon J. Barnett. (McGraw-Hill. N.Y. 1986). It contains a number of key programs reproduced in Cauzin SoftStrip form.

A post on the Classic Computers mailing list talks briefly about the Cauzin SoftStrip, but more interestingly, goes into detail about a greatly improved version of that technology that could put between 30-50k of data onto a 4x6" card.

You can supposedly still get software and service from Datastrip (Softstrip Inc.) Try this phone number: 203-573-0150.

Finally, many thanks go to Tony Cianfaglione (ab616@chebucto.ns.ca), Jonathan Galtry, Roy M. Silvernail (roy@sendai.scytale.com), and Charles P. Hobbs (transit@primenet.com) for providing information for this article, even though they didn't know they did. :-) Also, some pictures (SoftStrip Reader against a blue backdrop) are Copyright 1999 James Willing.

Software Trinkets

because

"Printed manuals aren't enough value."

History

While the late 70's saw the personal computer hardware revolution, the 1980's saw the computer software revolution, with big changes in the way entertainment software was produced and marketed. No longer distributed in zip-lock baggies, packaging was flashy and had professional illustrations. Software designers were glamorized, with their picture professionally photographed and placed on the back or inside of the box. Even the advertising blurbs were catchy and informally-written. Instead of a Xeroxed sheet of paper, documentation was professionally typeset--sometimes in color.

But perhaps the most interesting innovations were packaging add-ins: Additional materials other than the disk and manual that were included in a product. (Some of these "software trinkets" include the pink "wishstone" in Infocom's Wishbringer, the wizard's calling card and invitation in Transylvania, or black shiny moonstone and cloth map in Ultima 6.) The idea behind this new technique of marketing software--including additional materials in the box--was to give the customer the perception of added value. You weren't purchasing mere software, amounting to little more than a magnetic disk with a computer program on it; you were getting a complete entertainment experience. Early games didn't have real-time 3D graphics or sound, so any additional materials you received in the package helped immerse you more into the gaming environment. They helped make it real; they fleshed out the game's world enough to invade yours. (Hey, of course the magical land of Britannia doesn't exist--but it made it easier to pretend it did when you had a burnt parchment map of your surroundings.)

There is no question that Infocom was the first game company to pioneer this practice. Mike Dornbrook, who developed InvisiClues and had an illustrious career in Infocom's marketing department, recounts this history in the Interactive Fiction FAQ (part 2, to be specific):

There were actually quite a few people involved in creating the package elements for Infocom games. The game authors (we called them "the implementors") were the primary writers. The first exotic package was for Deadline (the third game, after Zork I and II). It was created because Marc Blank couldn't fit all the information he wanted to include into the 80K game size. Marc and the ad agency, Giardini/Russel (G/R), co-created the police dossier which included photos, interrogation reports, lab reports and pills found near the body. The result was phenomenally successful, and Infocom decided to make all subsequent packages truly special (a big benefit was the reduction in piracy, which was rampant at the time).

The first 16 packages were done in collaboration with G/R. David Haskell was the primary copywriter for Infocom materials (ads, catalogs, package elements, etc.). G/R typically did the "fluffier" pieces. Infocom's game implementor (and one of the co-founders) Dave Lebling wrote "The History of Magic" in Enchanter, but G/R wrote the "True Tales of Adventure" in Cutthroats.

We were spending a fortune on package design ($60,000 each on average in 1984 - just for design!), so we eventually decided to bring it in-house. I hired an Art Director, Carl Genatossio, a writer, a typesetting/layout person, and someone to manage all printing and purchasing of all the "feelies" in the packages. These folks (plus an occasional contractor during busy periods) did all the packages, hint books, New Zork Times, sell sheets, etc. from 1985 until the end in 1989. There were two writers during that time period - Elizabeth Langosy for most of it, then Marjorie Gove. Again there was a mix of game implementor writing and "marketing" writing. For instance, Steve Meretzky wrote the comic book in Leather Goddesses, but Elizabeth wrote the newspaper in Sherlock.

An unsung heroine of Infocom was our Production Manager, Angela Crews. She was responsible for acquiring the scratch-n-sniff cards, ancient Zorkmid coins, glow-in-the-dark stones, etc. which made the packages so distinctive. It was often an incredibly difficult task.

As for who oversaw all of this, again, there were many responsible. The Product Manager (first me, then Gayle Syska, then Rob Sears) worked with the Implementor and the Art Director to come up with a concept for the package and hammered out the details of the elements. All of these folks were intimately involved in the approvals, editing, tweaking, etc. which all of the elements underwent over a 3 to 4 month period. And many others (from the President, to Sales, to Testing) put in their two cents along the way.

I would estimate that each Infocom package had 1.5 man-years of effort invested in its creation.

FAYE: Who came up with all the gizmos in the packaging of the games (publishers, developers, etc)?

DAVE: The gizmos were originally suggested by our ad agency (which designed the first "interesting" package -- for Deadline). Marc Blank probably had something to do with it as well, as it was his game. After the incredible positive feedback we got, we felt we had to continue. The authors and the marketing department, in collaboration with the ad agency, came up with the goodies. There was always a question of what we could afford, and what was available in quantity, and so on. Once we were bought by Activision, there was additional pressure to lower costs, so the "feelies" (as we called them) became less and less wonderful. By the end they were most likely to be paper, rather than an object of some sort.

This particular Carny exhibit showcases some of the more interesting software trinkets to be included in PC games in the 1980s. By the way: I'm coining the term "software trinkets". If there is an official term for this kind of stuff (I know that Infocom called them "feelies", but I really don't like that term, since it isn't broad enough to cover everything here), please let me know and I'll use that term instead. Most trinkets were game-immersion items, like objects, artifacts, maps, etc; some were audio cassettes, either for instructional or entertainment purposes. A few rare trinkets were actually useful pieces of hardware, like a joystick or sound device.

Caveats: This list is by no means complete, and omits some of the common paper-only pack-ins, but will give you an idea of what I'm talking about. I encourage any readers out there to submit additional items for this list--some of the trinkets were pretty interesting. :-) Also note that any graphical links are JPEGs, and any audio links are either RealMedia or MPEG Audio Layer 3 ("MP3"). Finally, this list is in no particular order whatsoever.

"Type" Key:
   Immersion: Something used to help illustrate and/or define the gaming world you were entering.
   Cassette: Multimedia didn't exist back then, so audio cassettes were included for immersion purposes or instructional help.
   Peripheral: An actual thing that you could use to enhance the game (or other games). Not expressly limited to computer hardware.

Verdict

A brilliant idea. Packaging add-ins truly did provide value over the software itself--value you couldn't get from merely copying the software.

More Information

Paul David Doherty's "Infocom Fact Sheet" has pretty much everything you ever wanted to know about Infocom games, including a complete description and list of their packaging.

The Freeboard

because

"Home users require keyboard templates and hate cords."

History

As written above in "Software on ROM Cartridges", IBM designed and marketed a PC for the home user in 1984 called the PCjr. But unlike the PC and AT, IBM decided to create a new keyboard for the PCjr, with two drastic design changes: It was wireless, and it used rubber "chiclet" keys. They dubbed this keyboard the "Freeboard", no doubt a reference to its wireless capability.

The use of chiclet keys was an interesting decision, apparently made mainly to ease home users into computing. As Jeff Leyda recalls: "My copy of King's quest 1 had a template for the chiclet keyboard with all the commands on it. That was the whole idea behind the keyboard, so that software people can give you a template to put over your keyboard that shows what all the functions do." The wireless keyboard was also very convenient; you could type while reclining in your favorite easy chair, and since there was a RCA jack for composite color output to TV, you could hook it up to the family TV for everyone's enjoyment. While the keyboard was wireless, it could be connected to the PCjr directly for when you had no batteries.

Unfortunately, the PCjr didn't do well. The keyboard, what IBM thought would be considered the PCjr's most valuable asset, was ultimately its largest drawback. While it claimed you could be 6 feet or more away from the computer in wireless mode, most people found that, unless they had direct line-of-sight between the Freeboard and the PCjr, they had to be no further than about four feet from the computer. This distance limitation itself was only part of the problem, and is overshadowed by a more troubling philosophical issue: with a 14-inch monitor, you wouldn't want to get more than four feet away from the computer anyway--and if that was the case, what was the purpose of a wireless keyboard?

Another problem with the Freeboard was the size of the keycaps. In an effort to make the keyboard fit onto your lap easier, the key layout was compressed into a smaller form factor, so the function keys shared space with the normal 1-0 keys. To enter a function key, you had to hold down an "Fn" key as you pressed a number key. The numeric keypad was also omitted from the layout.

Finally, as if all this wasn't enough, the rubber "chiclet" keys proved almost unusable for touch typing, making word processing for experienced typists a chore.

IBM attempted some disaster recovery a year later with the release of a new Freeboard. Gone were the rubber chiclet keys, replaced by true keys (albeit still with a reduced-size layout). The power consumption and response of the keyboard's wireless operation was improved as well. But by this time, it was too late; other manufacturers had either eclipsed the PCjr's graphics or sound capabilities (Commodore 64, Amiga 500), or cost less (Commodore 64, Atari ST), or had a large established base of educational titles (Apple ][)--all of which appealed to home buyers more than the PCjr. IBM dropped the PCjr in 1985.

Epilogue: People today (1999, as of this writing) are starting to rediscover the practical uses of wireless keyboards. With the advent of video cards with embedded TV output, consumers are starting to hook their PCs up to the television, and a wireless keyboard is great for surfing the web and playing games while reclining on the family couch. Logitech came out with the first popular consumer wireless keyboard and mouse combination 13 years after the last PCjr was sold.

The Freeboard was a PCjr innovation way ahead of its time.

Verdict

Both a good (wireless) and bad (rubber keys) idea. Innovations aside, people don't spend $1000 to type on rubber keys.

More Information

The Oldskool Guide to the IBM PCjr lists a brief history and explanation of the PCjr, including all of its extra capabilities (and limitations).

Built-in Peripherals

because

"It sells more computers."

History

I found some PCs for sale at the local CompUSA today with built-in monitors (monitor, CPU, disk drives, etc. all encased in the same enclosure). I remarked out loud that while the iMac made this popular again, I never thought I would see a resurgence of it in the PC world. The salesman, a pimply-faced teenager about 12 years my junior, questioned me: "'Again?' PCs have never had built-in monitors; these machines are iMac copy-cats."

Never one to be called a liar, I thought I would bore him to death with a small, incomplete history of PCs with built-in peripherals. And, lucky you, you get to be subjected to this treatment as well. :-)

Not merely limited to built-in monitors, the first decade of the PC industry has seen other hardware embedded where no hardware should ever be embedded. I'm not talking about legitimate add-in cards and products, like joysticks, modems, etc.; I'm talking about big things, like monitors and printers. Things that probably should never have been embedded in the first place.

Monitors

Like the TRS-80, the Osborne, the Pet, the Lisa, and the Macintosh before it, "classic" PCs had a built-in monitor as late as 1987. They were mostly found in portable computers, usually called "luggables" because that's about how "portable" they were--you needed much muscle to lug them around. There was a single desktop that had an embedded monitor, however, and it wasn't from some crazy third-party vendor, as you would expect; it was from IBM itself.

The Compaq Portable was not only the first 100% compatible IBM PC clone (indeed, the very one that launched Compaq on its meteoric rise to the top), but the first PC specifically engineered to be portable. It was approximately XXX inches high by XXX inches wide by XXX inches deep, and the monitor and disk drives were all built into the casing. The keyboard attached to the front of the unit during transport, acting as a cover for the monitor and disk drives. The monitor itself was a green-screen monochrome monitor; but instead of having monochrome properties, it emulated the color CGA card, as CGA was more useful at the time. Colors were represented as shades of green.

The IBM Transportable was IBM's answer to the Compaq. It had very similar dimensions and properties. I seem to remember that memory expansion was more difficult on a Transportable than on a Compaq.

The Panasonic Sr. Partner was "a clone of a clone". It was a clone of the Compaq, but with additional features that, they felt, made it more attractive than an actual Compaq. It is mostly identical to a Compaq unit except it was roomier inside, which allowed them to build a printer into the unit (see below). The screen was monochrome-amber instead of green.

The IBM PS/2 Model 25 was the first (to my knowledge) PC desktop machine to fuse the computer and monitor together. Bundled with a space-saving keyboard (essentially a 101-key keyboard without the numeric keypad), the "8525" included a 12" VGA monitor, which could be either monochrome or color depending on how much money you wanted to spend. The Model 25 only had room for three slots (two if you had a hard drive in the second drive bay), so it wasn't meant to be a powerful expandable desktop; it was marketed primarily for students and educational use.

Printers

Yes, printers. Well, it made sense at the time: If you're going to be lugging your machine with you on the road, and you use it to do word processing, why not build a printer into the unit? It sure beats dragging a printer around with you.

The Panasonic Sr. Partner is the best example of a built-in printer on a PC. Located near the rear of the top of the unit, the thermal printer took thermal paper (a roll of fax paper, essentially) and had an effective printing quality of a 9-pin dot matrix printer. Since the Sr. Partner had a parallel port for hooking up to a conventional printer, a hot-key-combination was used to toggle the LPT1 device between the external port and internal printer. This toggle/redirection was done in hardware; no software was necessary.

Why a thermal printer, if the print quality is no better than a fax? Think about it; it made a lot of sense for a portable computer: Fax paper came on a roll. You could print 30 pages or more without having to (re)load paper, and the paper roll was stored inside the computer. This was very convenient for on-the-road printing. Also, thermal printers are very quiet (just like fax machines) when they print. The idea was that you could print out your report on the road without disturbing your traveling companions. Dot-matrix printers were, and continue to be, very loud and annoying.

Verdict

A bad idea. As we know today, you should never shoehorn your customer into proprietary configurations. Replacing/repairing built-in peripherals was difficult, if not impossible, and the extra mass outweighed (pun intended) the benefits.

More Information

I have scanned, converted, and made the Panasonic Sr. Partner's documentation available if you need it and a web search threw you to this page. It contains useful information in addition to the keyboard combinations for turning the printer redirection on and off. There are also some other web pages that mention the Sr. Partner, like McCain's Museum of Ancient Personal Computing, The IBM PC/XT Club, and Tom's Classic Computers..

Hardware Compression Boards

because

"Hard drives will always be expensive."

History

While not used much today, many people remember automatic compression products like Stacker, DoubleDisk, SuperStor, and Microsoft's own DoubleSpace/DriveSpace. Unlike utilities that required user intervention, like PKZIP or LHA, these software products allowed you to "double" the size of your hard drive by compressing the data before it got written to disk, and then decompressing it when you needed it. This all happened transparently; a memory-resident driver worked behind the scenes so the user wouldn't have to be bothered with explicit "compress" and "decompress" commands. Since the average compression ratio was close to 2:1, your hard drive's effective storage space was doubled. The only drawback was that the compression and decompression were done by your CPU, so your performance suffered; also, the device drivers took up anywhere from 20K to 45K of precious low DOS memory. But even with the performance hit and memory requirements, it's been estimated that millions of people used automatic compression software between 1990 and 1995.

But did you know that Stac and a few other companies came out with hardware compression products that did the same thing? Called "compression co-processors", these add-in boards had a custom chip that took the burden of compression and decompression off of the CPU. Two companies came out on top of this market: Stac, and InfoChip. Stac already had a history with hardware-assisted compression--their chips were used in many tape drives--so a software/hardware compression product for the PC seemed only natural for them. InfoChip, however, was a competing startup company that was created expressly for the compression market around 1990.

Hardware compression boards certainly offered benefits over their software-only equivalents. The major benefit was speed, of course: with a compression rate of 2.5MB per second (all specs used are from Stac's Co-processor cards), saving to disk became as quick as a normal uncompressed hard drive. But the real benefit was when you read files off of disk. At a decompression rate of 6MB per second, decompression became nearly instantaneous. This property meant that you could nearly double the transfer rate of your hard drive. This can be illustrated best with an example from 1991: Let's say you had a 2 megabyte file you were working with, and that your hard drive had a transfer rate of 500K per second. Reading this file into your editor would normally take 4 seconds (2 megabytes / 500K per second = 4 seconds). But if Stacker could compress that 2 megabyte file to 1 megabyte before it was saved to disk, and if the decompression of that file was instantaneous, that meant that your hard drive would only have to read 1 megabyte of actual data. Voila--your 4-second loading time just shrunk to 2 seconds. And this is on top of the fact that you are effectively doubling your storage capacity by simply using the product in the first place.

A minor benefit of the hardware boards was that the compression/decompression code could be moved out of the device driver, since the board was handling it. The device driver memory requirements then shrunk to as little as 14k (vs. a minimum of 26k for the software version) for Stac's product. But it was the speed, coupled with the fact that it could transparently compress the data on removable media (like floppy disks), that was the main thrust behind their marketing.

These benefits were indeed valid, even as late as 1993. But by 1995, 486s and Pentiums were fast enough that they could exceed the chip's maximum compression rate of 2.5 MB/s and decompression rate of 6 MB/s. Couple the speed of then-modern CPUs with 1995 hard drive prices, and you can see why the use of disk compression--even software solutions--had virtually disappeared. You could get an entirely new 500 megabyte drive for your $99, instead of buying the software version of Stacker. Or, don't pay any money at all--Microsoft had bundled a compression product free with DOS years earlier, which swiftly eliminated all other competition.

Here is a brief listing of the major hardware-assisted compression products that came out around 1990:

• Stacker's XT/8 Co-processor card. This was an 8-bit card designed for 8088/8086 PCs, and entered the market in 1990 with a street price of $149.
• Stacker's AT/16 Co-processor card. This was a 16-bit card designed for 80286 and higher PCs, and entered the market in 1991 with a street price of $199.
• InfoChip's Expanz! card. This was an 8-bit card designed for any PC, and entered the market with a street price of $199.
• Sigma Designs' Double-Up card. This was a MacIntosh product that used Stac's compression chip.
• Pinnacle Micro's Gold Card. This was a MacIntosh product that used InfoChip's compression chip.

Stac is now completely out of the compression business (at least for consumers; I'm unaware of how well their compression/decompression chip fares in commercial or hardware industries). As for InfoChip, they went bust sometime before 1994.

Verdict

A good idea--for about two years. After that point, hard drive prices fell hard and fast. Of course, they continue to do so today; as of this writing, the same $199 for Stacker 2.0's 16-bit board can get you an 8 gigabyte drive. (How's that for compression? :-) In fact, the only automatic disk compression product still around--hardware or software--is DriveSpace, which Microsoft automatically includes in Windows 95/98/NT.

More Information

A reprint of TidBITS from March, 1991 was one of the sources used in writing this section, as were reprints from June, 1990 and April, 1990. Additional information came from the comp.compression FAQ.

If you came here looking for technical help with Stacker, check out Stac's Stacker Tech Support Forum.

You emerge from the tent, dazed and confused. The bright sunlight seems to pierce directly through to your brain, giving you a headache... or was it the ozone inside the tent that made you feel sick? Or the dust? You shake if off and start walking back toward your car. Remembering your manners, you stop and turn around to thank the man for the exhibit. Reaching into your pocket, you pull out a couple of dollars and offer them to him, but he won't take it. "Telling their stories to someone willing to listen is payment enough," he says kindly.

As you get into your car and start the engine, he shouts back to you, "If you really want to repay me, tell others what you saw. Don't let these poor relics from an earlier age fade away into obscurity!" You weakly smile back, then drive off.

Later that evening, you call the resident "computer guy" at the office to ask him when the hamfest came into town. "Hamfest?" he asks. "There isn't any hamfest scheduled to come here until next month."

"But some guy out by Route 59 was showing me all this old computer gear," you reply. "I just assumed that the local hamfest was this weekend."

"Not that I know of. How much did he want for the stuff? I've been looking for a few things for my collection."

Come to think of it, he didn't name prices. "I'll call you back."

It takes 15 minutes to drive back out to the field--which is now completely empty. Nothing! No tent, no soap-box, no tall man. It's only been an hour since you were there, but nobody could've cleared out that fast... What the hell? What's really strange is that there are no tire tracks, no indentations in the grass--not even holes in the ground where the tent pegs were.

Only a lone BASIC cartridge rests on the ground where the PCjr used to be.