The following message was written by Mark Ryland in February 1991 in response to a message on PC Magnet's Power Programming forum extoling the virtues of Windows 3 (and Win32) and arguing that OS/2 was not necessary as an end-user environment.
##############################
> OS/2 is not necessary at the workstation level.
Wrong, wrong, wrong. Ask Microsoft -- it's scrambling to put most of OS/2's key features into Windows as we speak precisely because you're both wrong and yet tragically typical of PC consumers in your unthinking and unknowledgeable rejection of OS/2. Between people like you *and* the tricky problem of pushing Win3 while tacitly acknowledging that it's fundamentally half-baked, Microsoft has simply given up on the name "OS/2," and will sell you OS/2 next year (or the year after more likely) with the name "Windows" on the cover -- and you'll accept it happily, and indeed wonder how you lived with the current extremely crude and limited capabilities of Windows 3. What a gigantic waste of time and resources for you and the industry as a whole!
> I cannot see anyone developing a transaction processing system and
> not using some sort of database server such as SQL Server or
> Oracle as the storage facility. It is there that the fault-
> checking, stable multi-tasking, multi-user,etc. features are
> necessary, not at the workstation/terminal level. The 'client'
> should only be used to receive information from the server, and be
> used to develop requests for information/ updates to be sent to
> the server. Is there any other type of mission-critical
> application that will not be using the client-server
> relationship?
Look carefully at what you're saying. You're saying that end-users don't need fault-checking, stable, reliable computing systems. HUH??? What good is a server database if the clients are unreliable?
Every single user in a business organization should be using a multitasking system or else they're working inefficiently; and every multitasking system should be based on a fully-protected mode operating system, or else the result is: Windows 3, which is incredibly unstable, hard to configure, and unreliable -- and hard to develop for, which effects you as a user directly. EVERY COMPUTER IN EVERY ORGANIZATION IS SOMETIMES MISSION CRITICAL, and must be treated that way. If the receptionist in your company crashes his computer while recording a critical telephone call in e-mail, that could cost you far more than a routine crash of a "mission critical" database.
> As far as the user is concerned, new software should ideally be;
> o more intuitive to use.
Absolutely. (1) The number one technical requirement necessary to make a computer intuitive is *reliability.* Crashing and losing a day's work is the most *unintuitive* thing that computers can do. Here's a few other things: (2) easy to install and configure (ever had to muck around with Windows's SYSTEM.INI file just to get a network card to work in enhanced mode? or set VirtualHDIrq=NO to make Windows work with certain disks or disk managers? or endless configure QEMM and Windows to squeeze another 10k out of the DOS box?); (3) *actually* performs like it *appears* to perform (try explaining to your users about "cooperative multitasking" and why, although the system appears to be running programs simultaneously, they can't copy files to a floppy disk with FileManager or start Excel printing while doing file downloads under Windows); (4) allows for natural things like naming a file "angry msg to Lorne" <grin> rather than "AGYMSGLN.DOC"; (5) protects users from "dumb" but common mistakes (explain DOS's non-existent file locking (better yet, explain the workings of SHARE.EXE) to your Windows user who accidentally deleted USER.EXE while working on deadline).
> What I do take issue with, however, is whether the users - not the
> techies - need OS/2, or indeed, can handle OS/2 in its present
> form at this time. I think Win 3 is as big a step as they can
> handle over the next year or two.
Users need OS/2 MORE than do techies: Is "backing up" workstation files an intuitive notion, or an activity that users should be counted on to perform? of course not, it should be done over a network automatically without user intervention -- that's easy under OS/2 and nearly impossible under Windows. Is updating and distributing new application software to workstations simple and intuitive for users to do? Of course not, you spend valuable IS resources on doing it when it should be done over a network automatically and without user intervention -- again, easy under OS/2, impossible under Windows. What about sharing network resources? OS/2 transparently shares modems, printers, and disks/directories of both servers AND WORKSTATIONS -- try using a modem on the server with Crosstalk for Windows, much less a modem on another Windows client -- but that capability is built into OS/2, and *any* OS/2 comm program works perfectly with *any* shared modem on the net. Try explaining to your diskless workstation users why they can share files and printers on the network server with unmodified single-user software but not modems.
You seem to be asserting that OS/2 is somehow less intuitive to use than Windows. That simply isn't true, for the reasons just given and many, many more. Even looking at the issue superficially, the PM and Windows user interfaces are almost identical, and indeed the OS/2 File Manager permits a full icon mode with data files acquiring the icons of the programs that created them, unlike Windows. About all I can say in favor of your "Windows is more intuitive" position is that the Windows Program Manager is *prettier* than OS/2's equivalent, but it's not saying it's *more intuitive.*
> o more capable.
Agreed. Windows is a big step forward from DOS alone in capabilities, and OS/2 is a big step forward from DOS/Windows. And you want that -- look at how you're drooling over Win32 and using all this junk on a RISC workstation!
> o a relatively smooth migration path should exist, particularly
> where large investments in time and materials have been made (i.e.
> it should not be necessary to scrap all existing equipment and all
> prior experience with software each time an increment in
> capabilities is to be added).
OS/2 actually provided a very smooth upgrade path through the early emergence of character-mode OS/2 apps (WordPerfect, Lotus 3.0, Agenda, Rbase, Paradox, etc. etc.) -- unfortunately for all of us, that path was largely ignored by users. Now those same users are running the same character-mode DOS programs under Windows (a waste of resources, Desqview is better for this) when OS/2 would provide a far faster, more stable, and less memory-intensive solution (you know how much memory DOS apps take under Windows), and OS/2 can provide it all on a *286*, unlike Windows.
But that does lead to OS/2's single biggest weakness (from a marketing standpoint), which has been its weak DOS support. That problem will be solved with a vengeance in OS/2 2.0, which is the GREATEST DOS multitasker ever developed, far better than Win3. OS/2's other major weakness is lack of applications, but that's not nearly as serious as most people claim -- look at the sales figures of Windows versus WinApps. The former outsells the total of the latter by 5:1 or more; if people were using Windows *rationally* -- i.e., to multitasking WinApps rather than DOSApps -- the ratio should be 1:2 or more. That means that people are buying Windows and either (1) playing with it and then ignoring/deleting it (a common outcome); (2) finding the bundled miniapps adequate (an uncommon outcome, IMHO); or (3) running DOSApps on 386s -- the most common outcome, but not a particularly rational use of Windows.
> The most important thing about Windows 3 to me is, it allows me to
> get my department moving towards the things I think we should be
> heading towards, without waiting around for OS/2 to get its act
> together. MS's current strategy not only gets me moving NOW, but
> promises to keep the enhancements coming in the future.
* * * * *
> And in the meantime, MS is promising to make enhancements to solve
> most of the problems and bugs I can foresee (TrueImage will be
> REALLY handy for the local secretaries/bosses. multi-threading
> will be nice, 32-bits should make for some more interesting
> software EVENTUALLY, Win 4 finally promises to close the last gap
> between Mac & PC in user-friendliness, and OS/2 3.0 promises
> future RISC platforms - POWER! - plus a whole bunch more).
Fine, I think its fine to use Windows in the short term to get a taste of the real thing -- OS/2. But Microsoft is feeding you a line of sugar-coated BS. If you want computing power, as we all do, Windows 3 is the SLOWEST and MOST PAINFUL way to get to it.
1. Do you think that preemptive multitasking can be magically added to non-reentrant WinApps? Even when running on OS/2 with the SMK, Windows 3 apps STILL LOCK EACH OTHER OUT FROM THE CPU!!! They have too, otherwise they wouldn't be Win3 apps. Microsoft has already said that under Win32 current WinApps will be run as they are (all from one LDT and without preemption) and only new Win32 apps will run preemptively and with protected protect-mode memory operations (one LDT per process) -- exactly like running WinApps under OS/2 2.0. You'll have to get major new versions of all your Windows applications to take advantage of any of this, they will act and be treated differently by Win32 from your Win3 apps, and the total upgrade path for apps will probably occur in several painful stages stretching out over 2 years or more.
2. Do you think that threading can be easily added to WinApps? Not likely -- ask any OS/2 programmer, designing a program that efficiently uses threads often requires a complete rearchitecting of the entire application. A simple upgrade path for Windows developers? No way.
3. Do you think 32-bitness can be easily added to WinApps? Charles Petzold assures us that the 32-bit Windows API will be close to the current API, but that still means a major overhaul of all Win3 apps because the memory architecture (flat address space, no segments) is completely different. A painless upgrade path for Windows developers and Windows users? No way.
4. Do you think the High Performance File System can easily be substituted in for your FAT drives? No way. It's going to take a REFORMATTING of your drives, and all your disk utilities, caches, etc. are BROKEN, they DON'T WORK ANYMORE. Just like changing over to OS/2, except that you'll be starting several years behind OS/2 adopters in biting the bullet to get the tremendous advantages of HPFS. Oh, by the way, all your other DOS file-system kludges - LIKE NETWORKS, which are the ULTIMATE DOS file system kludge -- will break under Windows with HPFS, plus your CD-ROM drivers, your backup tape drivers, etc., etc., etc.
5. Do you think all these changes are going to come cheaply or easily or smoothly? Keep dreaming. You'll be paying Microsoft for a number of major Windows upgrades, and you'll be paying your applications vendors for a number of major upgrades. Do you think that the app vendors will upgrade you for free when they have to completely re-engineer their products to take advantage of Windows "enhancements coming in the future"? Lullaby, lullaby...
6. Do you think that a "Windows" system that supports HPFS, Named Pipes, Mailslots, preemptive multitasking, threading, etc. will run like Win3 with your current disk, CPU, and RAM resources? not to mention all this extra functionality will run sweetly on your DISKLESS WORKSTATIONS with ALL SWAPPING and IPC GOING OVER THE TINY 10 MBPS ETHERNET PIPE??? Rockabye, baby, on the treetops...
Conclusion: heavy investment in Windows will lead you down the most painful, most time-consuming, and most expensive path that you can map to the future: because the future of Windows is clearly a 32-bit fully protect-mode fully preemptive multitasking OS for every business user. Most of those features are *now shipping* in OS/2 1.3, and all will be shipping in OS/2 2.0 long before Windows32 sees the light of day. (Stop and think: since when did Microsoft or any other software company deliver a major product on time???).
Enough said. (Understatement of year, but you asked for it!!! <grin>)
