Sunday, May 24, 2020

Comdex 1999 by CryoModem - review of an early 80s text file

(This text file was written sometime in the 80's. I don't know who originally wrote it, but I've always enjoyed it. Now in 2020, I look back to see what was right and what was so very wrong...)

1999 COMDEX SHOW A FLOP

LAS VEGAS: The fall 1999 Comdex was, as always, a bit disappointing. The star of
the show was clearly the Yamagazi RoomTemp CryoModem I'm using to trasmit this
story. Yamagazi claims it blitzes out data faster than the speed of light, which
means that this report may have made it back to the office before the show even
took place.

(By 2000, consumer dial-up modems had peaked at 56 kilobit/s, with hardware compression on top of that to allow simple data to flow faster. DSL was on the rise although speeds topped out around 1 megabit both ways for commercial grade lines, and cable modems which promised faster speed were starting to roll out, but not as certain technology. BBSs and the concept of dialing a dedicated service were all but finished and the internet, while still young, had cemented itself as a necessary service.

In case it's not obvious, this paragraph is the whole premise of the article - that faster-then-light data transmission caused the report to travel back in time to the 80s.)

This year, the computer industry's companion show -Legaldex- nearly outdrew the
hardware and software exhibits. With so many pending lawsuits and so much money
at stake, it's really no surprise that Legaldex sprawled into 11 hotel lobbies,
two parking lots, and a hallway at the Liberace museum. the Apple booths alone
commanded more than 30,000 square feet of space.

(Lawsuits over software and hardware ownership and patents has never really gone away. Sun Microsystems was fresh in people's memories for suing over Java, but Apple still rings with an air of possibility. In fact, in 1999 Apple sued a PC manufacturer called eMachines, claiming their PCs looked too much like an iMac. eMachine's eOne was taken off the market as a result.

At the real Comdex 1999 - both Sun and Microsoft mentioned their mutual lawsuit in their keynotes - Gates with a joke and McNealy directly.)

In what has turned out to be an annual tradition, IBM once again trotted out a
new graphics standard, the 3DGA. Compatible with the MGA, MCGA, HGA, EGA, VGA,
EVGA, QGA, VVGA, VHGA, VQGA, VMGA, VAGA, EAGA, GAGA and, of course, CGA
boards, this new standard heralds a "bold new era of channel profitability," according
to IBM president and owner: "Now at last serious business users can have their fancy 3-D
graphs float in space."

(This didn't happen, at least with cards. After SVGA popular naming just sort of faded out, though there was briefly a UVGA. Of course, someone has to name everything, so the /resolutions/ still got names. By 2000 most machines were still 800x600 or lower, with 1024x768 possible but not fully supported by monitors. So based on that, the naming would have given us: CGA, QVGA, VGA, SVGA, UVGA and XGA. We didn't have widescreen yet, or at least not commonly. 3D displays were a long way off, with glasses or headsets still required most of the time even today. Glasses free 3D displays appeared around 2010, but didn't gain popularity.

Perhaps more importantly, despite creating the PC market and defining an architecture which survived and eventually defeated all comes, flourishing for  decades to come, IBM stopped being the driving force in the market in the 90's, eventually leaving it altogether in 2005 - though that's after this article. They had a brief resurgence in popularly in the late 90's with the Thinkpad laptop series, which was indeed a very good machine, but clone manufacturers dominated the desktop market and video card innovation was owned by dedicated video card manufacturers. In fact, the term "GPU" was coined by NVidia in 1999 and so would likely have been the graphical focus. 3DFX was the major leader at the time.

At the real Comdex 1999, 3DFX unveiled the Voodoo4 3D card and announced Voodoo5. the Voodoo4 would come with 32MB RAM, support AGP and PCI for $180. Voodoo 5 would come with 64MB or 128MB RAM, and cost $230-$600.)

Big Blue also displayed yet another new keyboard. The 143-key sports 6 randomly
scattered Ctrl keys, three more function keys, and an entire pad of SysRq keys
(though IBM did not annouce why anyone needs even one). To counter IBM's new Blu
architecture, AST/Quadram/Hyundai announced Blubus-Plus, with an additional data
line and slightly more shielding. Blubus throws off so much RF interference that
airborne users can make their planes bank left and right by leaning on the
cursor arrow keys.

("Big Blue" was IBM's nickname, a reference to their logo.

Keyboards didn't change much after the 101 key keyboard, although 104 keys became the standard after the addition of three Windows keys. Many keyboards also added media keys - some only a few and some a lot, but play/pause, next, previous, stop, volume up, volume down, and mute became relatively standard. SysRq seems to have gone away, though we still have pause/break doing nothing
most of the time...

New bus architectures did of course happen. By 2000 VESA had come and gone, and PCI was the dominant interface, though most motherboards still had an ISA slot or two for compatibility. PCI was released in '91 and caught on around '95, when Windows 95 introduced proper operating system support for it. For performance graphics, AGP was released in 1997.)

The fastest selling product at the show was IBM's just-released TBR (Technical
Bus Reference) manual, a fat compendium of IBM BIOS and chip-level errors that
the industry has had to accept as standards.

(Since IBM was no longer in charge, this didn't happen. Intel and Microsoft own the definition these days... although since IBM was still making machines in 1999, that may not have been the case then. I'm actually not sure!)

In response to the new line of IBM 240MHz machines, Compaq/Dell announced a
242MHz screamer, which it claims "makes the IBM box look like it's playing dead"
At the other end of the spectrum, we counted 35 manufacturers still selling
replacment motherboards for the original PC-1, switchable between 4.77MHz and
180MHz.

(In the mid-80s, where an 8MHz machine was considered serious and 25MHz insane, the idea of even a 180MHz upgrade board blew minds. Clock speeds in 1999 started in January at 450MHz and reached 600MHz  by the end of the year. Intel's Celeron was 400MHz, and AMD countered with their 450MHz K6-III offering.

This was the era of the Pentium 3, and there were no replacement motherboards for the PC-XT. In addition, the turbo switch disappeared during the 90's and computers just ran as fast as they could. This is also largely attributable to Windows 95 - a fixed feature set operating system meant that software could reliably use system timers, rather than CPU speed, to set their timing. The "Compaq/Dell" comment is interesting as maybe the only valid prediction in the document - if early! Dell acquired Compaq in 2002. Overclocking was big around the late 90s, though, so the 242MHz "Screamer" could have just been overclocked.)

Sponsors of next year's millennial Comdex are planning to call Comdex 2000
"Finally, the year of the LAN." Other vendors are proposing that Comdex 2000 be
dubbed "The Year of the Home Application," in an effort to prod the industry
into producing at least one product that could justify buying a computer for
home use.

(Hard to address such a tongue-in-cheek comment, but LANs were pretty established for businesses during the 90s, and the rising popularity of the internet was beginning to introduce them into the home - although it would really take the cable modem's victory and the spread of WiFi to fully integrate them years later.

As for the product... that's an ongoing thing. Arguably, though, the internet was the killer app that put a PC in every home. 20 years later, today, that's starting to fade a bit. Cell phones and tablets are replacing the general purpose PC for internet access.

As always, though, gaming also drives the PC market, as it did back then. 3D gaming was becoming big with 3DFX and NVidia pushing what the graphics card could do. AMD eventually replaced 3DFX as the big competitor.

At the real Comdex 1999, one author came away feeling that Sony's 64MB Memory Stick was the killer hardware of the show... perhaps it would have been if USB memory sticks hadn't followed on quickly, being cheaper and more compatible with the hardware that was out there.)

Ever-youthful Bill Gate's keynote address, "OS/9: The One You've Really, Really
Been Waiting For," blunted criticism that this newest version was still too hard
to use, too slow, and too memory-hungry: "Even though no third-party vendors
have taken taken advantage of the advanced capabilities of the seven previous
editions, dozens of developer are porting their applications over. And it will
run just fine on any system with 30 megabytes of RAM, although you may need a
bit more for your data."

(Bill Gates was only 44 in 1999. ;)

OS/2 failed out of the gate, and although IBM continued with OS/2 Warp in '94 and released the final version (Warp Server for e-Business) in '99, it only lasted a couple more years before being abandoned. Microsoft refocused on Windows with the release of Windows 95 and gained unassailable dominance for decades.

The criticisms remain, of course.

The porting comment was perhaps less an issue, the one thing that Windows did rather well was backwards compatibility. From Win95 onwards, applications generally "just worked".

30MB of RAM was laughable in the early 80s, but by 2000 systems came standard with 32MB-128MB of RAM, and virtual memory was standard, so the numbers, while big, were acceptable by then.

At the real Comdex 1999, a release date for Windows 2000 RC3 was announced and met with some doubt. Application compatibility was causing some delay. Windows 2000 was the first version to unify the NT and 95 kernels, so compatibility with both lines was important.

Bill Gates DID give a keynote, where he focused on Windows 2000, and tried to introduce the concept of the "personal web". It was seen overall as more of the same, so the article probably described it well enough.

Speaking of operating systems, there was a mini-expo for the Linux community at the same time. One writer noted: "while it's much nicer than last year's laughable bargain basement affair, it's small size gives ample evidence of the lengths Linux must go to enter the mainstream." However, Corel also showed "Corel Linux", a Debian distribution meant to be easy to set up.

Sun Microsystems also had a keynote, taking on Microsoft (with whom they were involved in an anti-trust suit), and pushed StarOffice heavily. CEO Scott McNealy's position was that software should be free (presumably hardware is the model?)

BeOS also had a booth that was well received, but there are few details beyond that. BeOS had its first x86 release in 98, but it was sold in 2001 and faded rapidly. A free reimplementation named Haiku was released in 2009 and was still active in 2018.)

In the word processing arena, MicroPro, Microsoft, and WordPerfect have packed
even more features into their bloated programs. MicroPro has purchased so many
third-party utilities that WordStar Professional Classic 7.3 is now delivered on
73 disks. WordPerfect has streamlined its 16-volume manual.

(Wordstar was abandoned in the early 90s as they failed to jump on the Windows bandwagon early enough. WordPerfect was also late coming to Windows, but held on due to a large install base. However, they were sold to Novell in '94, then Corel in '96. The 32-bit version of WordPerfect for Windows 95 was plagued with release issues and the final working version was rather late. By this time, Microsoft Office and in particular Word was gaining rapid market share. However, WordPerfect still survives today in 2020. I don't have a manual from 2000, but the WordPerfect manual today is only 282 pages long.)

Finally, Lotus announced its 1-2-3 WZ 50-dimension spreadsheet, a "quantum leap"
above its previous 1-2-3 VZ 5th-dimension version. Although users have been
demanding this added power, say market analysts, they're still not sure what to
do with more that three dimensions. When pressed for a delivery date, Lotus
officials would only say "Sometime in the first quarter of the coming millennium."
We can hardly wait.

(Lotus 1-2-3 was overtaken by Excel in the 90's for pretty much the same reason as the word processors - it failed to take the Windows update seriously in time. However, it did survive past the 1999 date here, eventually being discontinued in 2013.

As far as I can tell, 3 dimensions is as far as spreadsheets went, and even that is through multiple tabs rather than a single 3D sheet.)

... I spent WAY too long on this...


Sunday, March 22, 2020

Software is not a Science

I've been coming to a realization that one of the great problems with software development is the fervent belief that it can be simplified, that it is a science that can be nailed down to a fixed set of guidelines and will thereafter magically be perfect.

It's not. It's simply not. Like it or not, writing software is a creative aspect. You can no more reduce writing software to a set of fixed answers than you can create a checklist for drawing artwork.

Let's think about this.

First off, creating software requires the development of a unique solution to the problem from a set of incomplete tools which need to be assembled into a final system. It's a lot like building with Lego -- but you don't have the advanced set with all the fancy pieces. You have 2x2s and 2x4s and a couple of 2x8s.

It gets better. Normally the problem isn't even that wall defined. You have a Lego set but you don't have the instruction sheet and the box is torn so you only have half the picture.

Of course, we do have the continued advance of software development - new systems, new languages, new processes. These are all great and fancy things. These are Space Lego, and Harry Potter Lego. They let you more easily create the new worlds you are imagining. But they don't remove the creative element - you don't automatically get Hogwarts, you have to build it. That's how you bought Castle Greyskull, but not Hogwarts.

Most of the development processes that we develop fall into two categories.

The good ones aim towards giving us the instruction sheet. A set of processes that advance us towards our goal - with the understanding that it will help us build THIS product. We can use bits and pieces of one instruction sheet - creatively - to help build other products. But naturally you need to intelligently apply your creativity.

The bad processes aim towards removing the random creativity element, assuming that all problems can be solved in a fixed, predictable manner. Problem A is given to developer B who applies solution C, which takes time D. Neat, tidy, and guaranteed quality output. About as likely as Gingerbread men inviting you to tea and gumdrops, but we invest thousands and thousands of dollars in pursuit of this goal.

Software needs to be run more than art. Unfortunately, I don't have insight from professional artists on how it runs and whether it works for them. I'll need to do some research. :)