% fortune -ae paul murphy

From Chapter Three: The Windows Culture

This is the 27th excerpt from the second book in the Defen series: BIT: Business Information Technology: Foundations, Infrastructure, and Culture

Roots

In August of 1981, when IBM introduced its model 5150 PC:

• Companies like Tandy (Radio Shack) were selling tens of thousands of units per year -almost all of them running CP/M, often in dual processor configurations with the kernel running on a Z80C and user code running on a Motorola 68000 or later processor;

• The Apple IIe had a substantial installed base, a large stable of dedicated developers, and a successor - the Macintosh - in development;

• CP/M was the leading OS by volume and work had started on GEM, a full graphical user interface;

• Bill Gates II (the father) was a successful Seattle lawyer and an influential advisor to the IBM board of directors;

• Xerox was demonstrating an inability to succeed commercially with its Star 8010 - an advanced graphical workstation featuring a 17" greyscale screen; 40MB disk; and an applications suite for about $16.995;

• The University of California at Berkeley was making a big push for national prominence as a center of excellence in computer engineering; buying several Vaxes; hiring leading edge professors; and working with researchers from AT&T Bell-Labs to define sensible directions for operating system, language, and database research;

• Tandy's TRS-80, model 3, came with a 2MHZ Z80A and 48K of RAM in the personal model and an optional MC68000 running Unix at 4MHZ with up to 128K of RAM as a plug-in co-processor for business use;

• BIll Gates III, having abandoned his initial business plan (to build and sell traffic counters under the name Traf-O-Data) was focussed on selling Paul Allen's BASIC interpreter to hobbyists; and,

• Intel, whose 8086 chipset as introduced in 1978 was seen as too slow for high end computer makers and too expensive for low end manufacturers, was merely an also ran; mainly selling pre-assembled boards to hobbyists.

The Intel 8088, developed as a cheaper and slower version of the 8086, had been introduced in 1979 to meet demand from low end manufacturers for something that would work with older 8-bit peripherals, and was widely regarded as a dead end machine.

Thus the original IBM PC as introduced in August of 1981 was seriously behind the technology curve of the time:

• It ran slower than any of its major competitors - a full order of magnitude slower than high end machines from companies like Tandy/Radio Shack;

• It had essentially no software other than the command shell and BASIC interpreter;

• It had a stripped down operating system that had apparently been hacked out of a four year old version of CP/M and repackaged; and,

• It cost more than most competitors with a basic list price in the $1,995 to $3,000 range for a machine with one or two floppies, 64K to 128K, an OS license, a BASIC interpreter, a serial port, a parallel port, and a 12 or 13 inch black and white monitor.

  In contrast a complete Apple II system listed at $1,298.

Nevertheless, IBM sold about 13,000 units before the end of the year and nearly half a million during 1982.

Most of those units went to data center operators eager to experiment with any new offering from IBM. In most cases, however, the absence of a working COBOL compiler (the mid 1982 Microsoft COBOL for PC-DOS was not well received in the mainframe community) and any means of connecting them directly to the mainframe meant that most of the machines bought were simply shelved - many never even unpacked.

Some, however, went to commercial software developers to whom the access provided by the IBM name signified opportunity. In short order, therefore Wang ported its 3270 board and emulation to the PC, Satellite Software ported its WordPerfect 1.0, Dan Bricklin and Bob Frankston ported Visicalc to the PC, and Mitch Kapor ported his VisiPlot add-in software for Visicalc to the PC, added his own spreadsheet code, and sold the result as Lotus 1 2 3 for IBM.

By early 1984, when IBM introduced the PC/AT, three things had developed:

1. A clone building industry whose products were generally faster and cheaper than IBM's;

2. A software industry focussed on the IBM PC; and,

3. A magazine publishing industry focussed on serving advertisers in the PC marketplace.

The PC/AT's name and marketing program focused on the one thing it most clearly was not: Advanced Technology.

Released in December, 1984, the PC/AT had the 5.77 Mhz 80286 CPU Intel announced in February of 1982, five months after the original IBM PC debuted.

The core machine ran PC DOS and BASIC in 256K of memory but came with a remarkable list of options including:

• Customized 3270 emulation;

• Memory upgrades to 512K and 640K for use with PC-DOS

• Memory upgrades to 3MB for use with Xenix, a Microsoft Unix variant;

• A network card for use with Xenix and the then unannounced IBM Lan Manager product for PC-DOS;

• An Intel 80287 arithmetic co-processor; and,

• A board with two MC68010 CPUs that ran native 370 code, including VM, DOS/VSE, and MVS within the PC/AT architecture.

For most people, however, the options that counted were:

Component	List Price (January, 1985)
Base PC/AT, 256K, 6Mhz 80286, 1.2MB Floppy (no hard disk, monitor or graphics board)	$3,995
Fully functional base package - 10MB Disk, 256K, 1.2Mb Floppy, base monitor and keyboard, PC-DOS licence	$5,500
512K RAM upgrade (to 640K for dos users - a board swap from one 256K board to one 512K board plus one 128k add-in.	$1,125
Minimal graphics adapter (EGA card)	$542
Basic color display for EGA card (15")	$849
IBM's PC-DOS	$40
Floor standing enclosure	$165
Intel 80287 Floating Point co-processor	$375

Here's how a reviewer in early 1985 described the PC/AT's graphics options:

More appealing to the mass market is the new Enhanced Color Graphics Display ($849) and the Enhanced Color Graphics Adapter ($524 to $982 depending on options). The adapter can be used to drive any IBM display. On the IBM monochrome display it can generate the current high quality text (characters are 9 by 14 pixels) as well as graphics with a resolution of 640 by 350. On the IBM Color Display (model 5153) or compatible RGB monitor it can duplicate all of the standard color graphics modes, but also delivers 16 colors to both medium-and high-resolution graphics. On the new Enhanced Display (model 5154) it can duplicate the old modes and can produce images with a resolution of 640 by 350 pixels using 4 to 16 colors.

The most stunning thing about this machine was the market reaction to it. Over the next year something like 1.4 million units were sold - mostly at list prices to people who had other choices including:

1. The multi-user Tandy 6000HD owned the low end business market . At $4,499 this offered 512K of RAM, a 15MB hard disk, a 1.2MB eight inch floppy, an integrated 24 x 80 screen, and two CPUs - a Z80C for backwards compatibility with TRS-DOS and a 7.54MHz MC68000 [with 32 bit data registers and 16/24 bit addressing] that ran Xenix off the hard disk.

   The Tandy offered lower cost, had national distribution, a deep installed base, and included a wide range of business applications covering entry level financial management, word processing, and even automated UUCP (the internet's predecessor) network access.

2. The Sun 3/160 with a 16.67MHZ MC68020 full 32bit processor owned the high end design market. WIth 16MB of RAM, a 21 inch greyscale monitor running at 900 x 1152 pixels, a 60MB disk, and an integrated ethernet port this cost about $18,900.

   The version roughly as shown below, with dual 21" monitors, a 20Mhz processor including the MC68881 math co-processor, 32MB of RAM, and dual internal 104MB disks as well an external SCSI controller cost about $38,800.

   The Sun 160 offered engineers and others working with high end graphics or other software an open, network based, environment with considerably more horsepower than a low end VAX at a fraction of the latter's cost.

   Relatively few were sold, but the machine was a major hit among the academic researchers whose work established the CAD/CAM (Computer Assisted Design, Computer Assisted Manufacturing) industry that was to the power the enormous productivity gains in the US economy in the mid to late nineties.

3. The Compag DeskPro was a better IBM PC/XT Based on the 7.14MHz 8086, it cost about half of what the AT did, and was generally faster when running existing applications.

   Compaq was also the first to clone the AT, producing its substantially cheaper 286e by mid 1985 and was the second company (after ALR) to transition to the 386 the following year. Because Compaq had written the hard disk extensions to MS-DOS 2.0 (making it 2.1) while the PC/AT shipped with the somewhat buggy and often incompatible PC-DOS 3.0, Compaq's ads at the time call their machine more compatible with IBM, than IBM's own AT.

4. The MacXL defined its own genre. At a list price of $5,495 ($5 less than the base PC/AT) this offered 1MB of RAM on a MC68000 at 7.54Mhz (5 on early models) with a 10MB disk, a 720K 3.5 inch floppy, the MacOS GUI and a full suite of graphical applications on a high contrast, black on white, 640 by 480 screen. About 30,000 (1%) of PC buyers picked this unit.

   The Macintosh offered the home and professional user a complete set of graphical applications with all the advantages of Postscript (see note below) printing for less money than the "bare metal" PC/AT.

   It became a major hit with the printing and publishing industry because of this, but the defining characteristic of companies buying it was that they had no formal systems departments in place prior to getting their first Macs.

In all four competing types of machine the software worked out-of-the box and the hardware was generally reliable within the context of the day. Neither statement was true for the AT, mainly because Microsoft's adaptation of MS-DOS for the 8086 did not allow smooth migration to the 80286 environment and a majority of third party board makers didn't initially get full access to the software APIs.

Here's what an early reviewer not working for a PC dominated magazine found:

Programs written in Basic for the PC ran without a hitch, which makes sense since Basic 3.0 bundled with the AT contains only minor differences from Basic 2.0 bundled with the PC. Many business packages, including word processing and graphics programs, also worked without a hitch.

One curious result of our testing was an "insufficient memory" message when we tried to install WordStar. Our AT had 640K RAM, yet the installation program refused to believe that it was there. We never did get WordStar to work, but we learned that WordPerfect works quite well.

Of course, the biggest test of PC compatibility is running Lotus 1-2-3 and Flight Simulator. Unfortunately, the AT ran neither. We could start Flight Simulator, but the program soon froze.

All in all, a little more than half of the PC software we tried on the AT ran without problems. Some loaded fine, but bombed out during operation. Some would not load at all. Our best advice is to try your favorite software first.

Reviewers working for PC Magazine, Byte, and other publications dedicated to the PC told a very different story. To them the PC AT was a magnificent technical breakthrough (it wasn't) costing less than the MacXL (it didn't), running faster than anything else (it didn't) offering 32bit computing (it didn't), breaking the graphics mold (it didn't) and running existing software failure free (it didn't.)

What it did do, however, was establish Microsoft's operating system software as the industry standard. It is hypothesized that unrestricted application software copying in the MS-DOS environment had a role in this since office software purchased from IBM for PC-DOS (IBM's name for MS-DOS) would not work on a Mac or under CP/M86 but the evidence for and against this is equivocal.

It is clear that mainframe management's control of corporate systems budgets and their commitment to IBM influenced purchasing and this clearly created a something-for-nothing opportunity for a lot of people willing to take home office software, but how important this actually was in building its market share, is unknown.

Some notes:

1. These excerpts don't (usually) include footnotes and most illustrations have been dropped as simply too hard to insert correctly. (The wordpress html "editor" as used here enables a limited html subset and is implemented to force frustrations like the CPM line delimiters from MS-DOS).

2. The feedback I'm looking for is what you guys do best: call me on mistakes, add thoughts/corrections on stuff I've missed or gotten wrong, and generally help make the thing better.

   Notice that getting the facts right is particularly important for BIT - and that the length of the thing plus the complexity of the terminology and ideas introduced suggest that any explanatory anecdotes anyone may want to contribute could be valuable.

3. When I make changes suggested in the comments, I make those changes only in the original, not in the excerpts reproduced here.