I often end up describing the state of the computer industry in automotive terms. I would say that in many regards, the computer industry is much where the auto industry was in the 1950s and 1960s. Lots of flash, lots of glitz. New models every year. Lotsa color. Incredible options. Not so good at starting reliably. Prone to leaving you stranded when you really need it.
This is much where the computer industry is today. People line up to see and buy the newest and fastest computers. They line up to buy the newest software. And yet, if asked, they would normally acknowledge that no, they really don't expect the product to work any better than last years product.
Contrary to popular opinion, the reason the auto industry build beautiful, unreliable cars in the 1950s and 1960s wasn't that Detroit was evil, but that is what sold. It costs money to build a high-quality car, and if the consumer can't be induced to pick up the cost, it won't happen. Also, incredible lists of optional features are contrary to high-quality construction. If you look at the list of options, and thus, combinations of options available in cars of that era, it is incredible that the cars could even be assembled. Many cars had a list of five or six engines, several transmissions, several rear-ends. Widely varying interiors, trim options, etc. YES, lots of options still exist on cars today, but most are in the form of "packages" -- you won't be able to order the big engine with the little radiator. On top of this, customers expected radical changes every year to PROVE to their neighbors that they had the NEWEST version of the car, not a three-year old model. Radical changes every model make manufacturing a quality product very difficult.
Attempts to manufacture cars with the requisit simplicity and consistance to build a reliable product didn't work. Until the late 1970s and 1980s, the Japanese and German cars had but a small nitch in the market. Quality built, boring cars rotted on the lots. One can't even blame the Auto Industry for not following the market immediately when the market started to shift: Only a few years before the 1978 oil crunch, the 1973 Arab Oil Embargo raised interest in small cars, but only for a VERY short period of time -- the only American manufacturer to radicly down-size their entire fleet was American Motors, and they paid dearly for that "mistake" when the price of gas fell and interest in small cars vanished.
Oh, yeah...this is supposed to be about the computer industry, isn't it??
Anyway... Fashionable as it may be to bash Microsoft, in reality, they are making what sells. If *WE* didn't buy Microsoft products, THEY wouldn't be making them! If we start refusing to buy an update with just more bugs and more problems, guess what? It will change. Right now, a company which ships a product out first is rewarded by market share. A company which holds back their product because they think there are problems to resolve gets punished by loss of sales to the competitor with the "bad" product.
This isn't new. Years ago (1986 or 1987), when the very first 386 machines were getting ready to ship, Intel sent out a warning to all the manufacturers who had purchased chips that it turned out the first batch of 386 chips were missing some thing: the 32 bit multiply instruction. Something happened between prototype and production, if I recall, the multiply hardware was just plain missing. I worked for Zenith Data Systems at that time -- when word of this bug was discovered, our factory (according to sources who should know) had the machines sitting on the shipping dock ready to go, but they were held. A major competitor of ours, Compaq, went ahead and SHIPPED this KNOWN DEFECTIVE product. Now, at the time, Compaq was a manufacturer of incredibly boring, not innovative desktops, and huge machines they called portables (which actualy gave the company its name). As I recall it, they were not in any particularly great financial shape, and in fact, as I recall it, they were rather unhealthy. This machine changed their image from boring to exciting, cutting edge stuff, in spite of the fact that what they shipped was KNOWN DEFECTIVE. By the time software came out that USED the new 386 features like 32 bit math, the first Compaqs were well out of warranty, and besides, no one cared -- most of these machines were too slow to run it!
This is my version of history. I was there. I've done no research, however, to verify my memory of events. If I had a PhD, you could quote me very safely on this. Since I don't, you had better double check it before you do.
I haven't liked Compaq since, although they have given me plenty of additional reasons... They managed to commit fraud and the market rewarded them. No wonder we keep seeing more of this type of action
Another hardware example: US Robotics (now owned by 3Com, which
really trashed their name in my mind): In the early and mid-1980s,
they were a minor manufacturer of modems. They then shipped a very
bad modem called the Courier 2400, one of the the first 2400 bps modems.
It used a known defective chip from Rocwell which worked great when talking
to other Couriers, but not to any other 1200 bps modems (the best most
people had), and particularly poorly talking to US Robotics 1200 bps modems
(I had one!). They pulled a coupe by arranging special ultra-low
cost purchases for bulletin board operators (I was one at the time, I didn't
do it, however): If I recall properly, the price was something like $300
which was cheaper than most people paid for 1200 bps modems at that time.
The result: Most BBS operators switched to the USR modem. Most users
had to upgrade their modem, or live with 300bps (for those who think 28.8kbps
is slow!). To be fair, USR did a pretty good job of fixing the defective
modems, but still it ended up being sell-first, engineer-later. Not
cool -- in my book. However, USR became THE standard of the modem
industry, in spite of the fact they have pulled the same basic stunt time
and time again. Hayes, the previous "standard" modem manufacturer
started its downward slide here, for they waited until they had a VERY
GOOD product to ship, rather than shipping the first thing that actually
managed the higher speed connection.
I recently thought of another example... Imagine a room full of
people. Imagine a hand grenade being tossed into the room.
Now, in the real world, perhaps one selfless soul might throw themself
on the grenade to protect and save the lives of all the others, knowing
full well that they face certain doom. Most likely, however, everyone
will hope they will be a surviver and perhaps someone else will jump on
the grenade. Now, if instead of normal people, we have a room full
of computer software buyers, and someone tosses in a grenade such as an
upgrade of Windows 98 or Internet Explorer 5.0 Beta, they will be falling
all over each other to toss themselves on the grenade. Bizzare.
When Windows 95 came out, the question I received is "When can we upgrade?" Now, with Windows 98, the question I normally get is "Do we HAVE to upgrade?"
About time. The solution isn't in Microsoft. The solution
is in the buyers. You can lead a buyer to market, you cant make them
buy a quality product. (something tells me that won't become a new
popular saying...)
(c)opyright 1998, Nick Holland