Re: www.microsoft.com sure needs a lot of silicon

On Sat, 26 Apr 1997, Erik Funkenbusch wrote:

> Tracy R. Reed wrote in article <5jpd3o$ao7$2_at_hole.sdsu.edu>...
> >Srinivas Thimmiah (tsreddy_at_netcom.com) wrote:
> >: You can burn me at the stake for heresy, but perhaps the server field
> >: is now ready to standardize on ONE version of OS (just like the PC
> >: industry standardized on MS/DOS) for all the advantages of economies
> >
> >The PC industry never standardized on MS-DOS willingly. MS forced it on
> >the industry. We literally had no choice. Novell and several other
> >companies tried to introduce their own DOS's, and MS crippled their
> >software to recognize those other DOS's and their software refused to
> >work under it. MS also sued Novell and perhaps someone else over the DOS
> >issue. The PC industry has also standardized on Windows, for very much
> >the same reasons.

Correction, IBM used it's marketing muscle to force MS-DOS onto the industry. IBM wanted to make sure that the PC never got beyond a "FAT Terminal" in terms of paradigm. The PC was supposed function as a terminal in which a "form, document, or spreadsheet" could be uploaded to the mainframe and processed in batch.

MP/M had already killed the Series 1/EDX market (just as IBM planned to make a "killing" with it's EDX 3.0 release) in 1981. IBM talked to Gary Kildall, who was preparing a multitasking, multiuser implementation of MP/M which IBM wanted to prevent. In a PANIC they had Gates provide MS-DOS (which was purchased from a known plagerism of CP/M, but there were no "deep pockets" this way).

> Oh give me a break. OS/2 was out in full force prior to Windows 3.0's
> introduction. OS/2 had the lead, why couldn't they keep it?

IBM had deliberately prevented Microsoft from implementing mulitasking and multiuser capabilities that were desparately needed to compete with UNIX.

> It wasn't as
> if everyone just jumped on Windows 3.0 the second it was released. It
> wasn't until 3.1 came out that any real applications started to hit the
> market (almost 2 years later and just prior to the release of OS/2 2.0).

Actually, Microsoft promoted the daylights out of Windows 3.0. The main advantage is that it "could run" under a 8086, 80286, or 80386. It was practically brain dead in all but 386 mode. Microsoft had tried unsuccessfully to promote windows 286 and windows 386, both of which were little more than "software carousels". In 1990, Sun came out with the IPC lunchbox, directly targeting the high end Windows 3.0 market. The price of the IPC and the price of a Windows 3.0 capable box were within a few hundred dollars of each other. The IPC would be shown running 20 strip charts, bouncing balls, scrolling headlines, and real e-mail and netnews.

By July of 1991, Gates realized that he had to compete head-to-head with Sun, and win. It was at that point that NT was first announced, that Microsoft cut it's ties with IBM, and that OS/2 2.0 became the "astable release of the week", many OS/2 users refused to give up OS/2 1.2 until the release of OS/2 WARP, and even then only reluctantly because of the lack of grandfather support. McGraw-Hill spend 3 weeks trying to scrounge up a legitimate copy of OS2 1.2 for a corrupted server. It seems someone had put a magnet too close to one of the floppies.

> In other words, there *WAS* an alternative before Windows became a
> standard, but most people didn't want it.

Actually, people wanted Sun. Unfortunately, Sun dropped support for it's Intel implementation of SunOS (sun 386) and put the collection into general public license (which became the foundation for the Linux application suite) on SunSite.

> >: of scale and the only company which is delivering a solution (however
> >: bad it might be) is Microslop.

McNealy made 3 major mistakes (in hindsight):

It didn't push for an
"under $300 Unix License" from AT&T. AT&T had set a "floor" of $700 for UNIX back when a PC running MS-DOS cost over $5000 (1983). When Novel purchase the Unix Rights in 1992, they shot themselves in the foot by refusing to include TCP/IP and NFS. By 1992, Linux had developed a small grass roots support.

Second, they tried to shut out the Intel market prematurely. The down side of the Intel market was that there were hundreds of different drivers and thousands of possible permutations. The configuration problems were a nightmere. The only possible way to get the needed support would be to put some of the driver source code out there so that users could create and link their own drivers.

Finally, Sun alienated the rest of the Unix Consortium, creating rival OSF/1 and encouraging customers to "hold out" for OSF/1. By the time OSF/1 arrived (1992), Microsoft was announcing that Windows NT, a "better Unix than Unix" would be out "Any Day Now". This "beta release pending continued for all of 1993 and most of 1994, Microsoft was rumored to have even used an SCO kernel at one point just to create a working "press release demo".

Nobody was happy with Windows 3.1, it crashed several times/day, the hourglass was a constant companion, and the 8.3 file names were getting a bit tedious, as was the 512meg drive limits. 2 liner .bat scripts were eating 32k of disk space. Attempts to compress, repartition, and otherwise manage the unmanagable was wearing thin on the patience of managers and staff alike. Microsoft couldn't even keep it's own applications alive.

In the period from 1992 to 1994 Microsoft spend over $1 Billion/year in full page advertizing to keep the press from pointing out that "the emperor has no clothes".

When NT was finally released, it required substantial amounts of memory, was very slow, and the user interface was actually worse than windows 3.1. Even worse, the configuration was hellish. You had to be intimately familiar with hardware to be able to install the stuff. Attempts to guess at which hardware belonged to which interrupts and which memory buffers were nearly hopeless. The Windows 3.1 "thunk mode emulation" slowed the machine to a crawl even with 32 meg of ram. In terms of cost/benefit, it was more expensive than a comparable Sun SparcStation or IBM RS/6000.

> >Because they are the only ones capable of delivering a solution.
They aren't delivering a solution. The NT 4.0 server is being reluctantly adopted. The windows-95 workstation is being tolerated until the promised release of Windows 97. The standard joke is "Back up the registry 20 times and pray". The Microsoft version of the Rosary, for each hail Mary, add an application. For each Our Father, back up the registry, or is that the other way around?

> >Anyone
> >else is run into the ground through hostile business practices. Are you
> >really saying that *no* other company in the world is capable of coming
> >up with an OS which is better than Windows? We all know that's gotta be
> >bogus.

The saga of Windows NT/95 is like the parable of the Emperor's New Clothes. No one dare admit they couldn't see the cloth for fear that they be considered unfit for their job. It's only when the town fool or the youngest child, who has no station whatsoever, hollers out "The king is as naked as a jay-bird" that people are finally willing to admit what was obvious.

Microsoft holds 9 billion dollars in equity interest in several hundred major corporations. Enough to put the stock of any single company or group of companies into a major skid. Much of the holdings are in media companies. In addition, Microsoft has generated "Partnerships" through companies who misunderstood the terms of NT Server licenses. The net result is that no Senior Manager can admit that Microsoft hasn't delivered what it has promised for 5 years now "A better Unix than Unix".

Meanwhile, like the mammals in the ice age, Linux users have been quietly proliferating, putting linux partitions on their workstations and laptops in such a way as to make it invisible, except to themselves. The only clue is the telltale "LILO" message that displays for 5 seconds after the system boots. The desktop looks similar enough that a Linux user can be working with what appears to be a Windows 95 workstation and is actually a Linux workstation. Giveaways include the "9 panel selector" and the performance meter scrolling at the bottom of the screen. The contributor chose to remain anonymous.

Many middle managers are generating small coups, by allowing their linux users to configure a linux based "Intranet Server". By the time the procurement people get through the process of approving an NT Server and 200 user licence pack array, the Linux server has been up for months. It took longer to write the RFP than it took to load Linux on an old 486 that in been discarded in favor of a Pentium 120 required for a Windows NT or 95 upgrade.

> Better does not equal successful. I won't start another Beta vs. VHS
> argument but it's been proven.

I was there, selling VCRs when the first Beta and VHS machines came out. Beta was really nice, it had great special effects, very precise queuing, and all the features a video editor would consider critical in quickly editing a series of video clips.

Over time however, the U-Load would pull the tape out over the head, crinkle the tape, and eventually, after dozens of reloads at the exact same location (the start of the tape), would trash the video tape.

Unix was never terribly user friendly. It took an MSEE to install and a BSEE to maintain. Even the night operator had to be a BSCS. A qualified Unix administrator makes a six figure income in the northeast. Only recently has the NT administrator pay reached that limit.

After successive rollouts of NT 3.51, Windows 95, and NT 4.0 server, Corporate America isn't exactly rushing into another round of $5000/user NT 4.0 rollouts. As they become more aware of the possibility of $1 Million dollar NT server License Pack royalties, middle managers are becoming more interested in alternatives.

Linux is just a "babe in the woods". Speculation about the tactics of Linus Torvalds after his graduation (Would he take ownership of Linux?), and the very gradual growth of tightly held companies like Red Hat and Info-Magic made it very hard to get a clear picture of what was really happening in the Linux Market. A sudden burst of Red Hat 4.0 and 4.1 releases, generated by a customer base of highly satisfied customers, on a product which has never been advertized outside of the internet (the advertizing budget for Linux was less than $5 Million), and overwhelming response to Linux Celebreties (Linus Torvalds and Patric Volkerding) have caught all but current Linux users completely off guard.

What we have seen in the last three months is shocking. Not only has Microsoft not delivered Windows 97, they have announced that NT 5.0 will have many of the features currently provided in Red Hat 4.0. Meanwhile, 3rd party software vendors, hungry for an alternative to the NT marketplace, dissappointed in the Mac and OS/2 marketplace, and looking beyond year 2000 in terms of hardware and software, are suddenly discovering that commercial application software for Linux is actually selling quite well. Linux may be the hottest "boom" since the "internet rush" that hit when Netscape went Public.

I'm sure Microsoft will do everything it can to prevent coverage and minimize the damage. Unfortunately, it is very hard to argue with 6 million satisfied customers on all 7 continents. Linux is even available on DEC Alpha, Sun SPARC, Power PC (Mac and RS/6000), and MIPS (Silicon Graphics). Your applications are available in binary and source form, and the commercial software will likely be available for all platforms as well.

Most importantly, Applix and Corel are both offering Linux products that can Import and Export Microsoft Word and Excel documents (Yes virginia, Mr Gates will even get Royalties on Linux). Received on Wed May 07 1997 - 00:00:00 CDT