The following letter was posted on the Programmer's Connection BBS. As per the author's request I am taking it upon myself to post it in CompuServe to give it as wide a distribution as possible. I too believe that Mr. Miller's article was a heavily biased rip-job aimed at OS/2. It should not be allowed to stand without rebuttal. Steven Kahn [76167,140] April 27, 1992 Message : 11335 (OS/2 Version 2 (Beta Test) ) Date... : 4/12/92 (6:23 Pm) From... : Richard Hodges Refer.. : None To..... : All Sec'ty. : Public Subject : Open Letter: PC Mag Rec'vd. : No Anyone/Everyone: As I mentioned earlier, the PC Magazine article by Michael J. Miller has gotten me 'riled up'. I have written an Open Letter to PC Mag which appears in the next 8 messages. (For those of you who don't care to read it, just skip 8 messages ahead.) I am sending a copy to PC Magazine, but due to the length of the letter and its pointed criticism of an article written by the editor-in-chief it is clearly impossible that this will be published. Nevertheless, it would be worthwhile to have an opposing viewpoint published somewhere. The only option appears to be electronic distribution. If you agree with the thrust of my message and would like to see this message distributed to others, please capture the letter and cross-post it on other bulletin boards and information services. For example, I am not on Prodigy, CompuServe, BIX, or PC Magnet - it would be good to place it in public forums there. If you like, print it out and send it around your department at work! Also, WRITE to PC Magazine. If your feel strongly about some particular point, let them know about it. PC Mag needs to know two things: 1. There are plenty of people interested in OS/2. 2. We aren't going to sit quietly and let the press kill OS/2. Thanks, Richard Hodges ---------------------------------------------------------------------------- AN OPEN LETTER TO PC MAGAZINE Richard E. Hodges University of California, Los Angeles and Jet Propulsion Laboratory Disclaimer: These are my personal views and do not reflect the official position of my employer. I am not an employee, consultant or stockholder of IBM Corporation. I do not have any financial interest in DOS, Windows or OS/2, or in any applications programs designed for use on these systems. INTRODUCTION I have just read the April 29 issue of PC Magazine and strongly disagree with the observations and conclusions of Michael J. Miller in comparing Windows and OS/2. This letter provides a point-by-point rebuttal to the main points that I have found objectionable in two articles by Mr. Miller. This will be addressed below. In contrast, the article "OS/2 2.0: Does It Fulfill the Promise" by Joe Salemi provides a fair and honest appraisal of OS/2 2.0. Joe Salemi has done a masterful job learning the facts and presenting a useful overview of OS/2. His feature article and additional sidebars, "Guided Tour of the Workplace Shell" and "Debunking the Myths of OS/2" are refreshing in that they directly report the facts without introducing personal preferences and bias. It is indeed rare to find a journalist that maintains his objectivity when evaluating operating system software. Another highlight of this issue is the short piece by William F. Zachmann, "32-bit GUI Alternatives." Mr. Zachmann has clearly and accurately delineated the key reasons that OS/2 is more appropriate on the desktop than any currently available versions of Unix. As a person that uses Unix based systems frequently, I can personally confirm that Mr. Zachmann's evaluation is correct. It is useful for this information to be summarized for those who may have heard of Unix but do not have first hand experience with the system. Finally, Charles Petzold's piece on Windows NT holds some academic interest, but seems inappropriate in an issue devoted to currently available desktop operating systems. This article is particularly inappropriate in view of Microsoft's statement that NT is a server operating system, not a desktop system. Furthermore, Microsoft's abysmal track record on meeting an announced delivery date with systems software causes one to seriously question the credibility of statements such as, "Microsoft is predicting the retail release by the end of 1992." This credibility is stretched even further by the fact that to this date, Microsoft has not yet produced a reliable multitasking operating system. WINDOWS ENTERS ITS PRIME; OS/2 IS A FASCINATING TECHNOLOGY BUT A DUBIOUS CHALLENGER My complaint with Michael J. Millers article is primarily centered on his general recommendation of Windows 3.1 over OS/2 2.0. I can appreciate that Windows 3.1 is useful to people with relatively modest requirements or low end systems which cannot run OS/2. However, for those of us who have more capable hardware, there is no need to be stuck with the limitations of Windows. Thus, rather than make a blanket recommendation as Mr. Miller does, it would be more appropriate to recommend the system that is best for a particular class of hardware. It appears that Mr. Miller feels compelled to try to justify his dubious recommendation. In what follows, I will identify some comments that appear to be designed to support a prejudiced conclusion. I wonder why Mr. Miller says, "... developers can write 32-bit applications, which theoretically can be faster than 16-bit applications normally created for Windows or OS/2 1.x." THEORETICALLY? This is not a theoretical hypothesis: it has already been proven with 32-bit DOS extenders and with beta versions of 32-bit OS/2 applications that 32-bit flat memory model programs are from 50% to 100% faster than 16-bit segmented memory model programs. One would expect the editor-in-chief of PC Magazine to be well aware of this fact. From the context of his statement, it appears that Mr. Miller is attempting cast some shade of doubt on the established fact that 32-bit applications run faster and thereby cast doubt on the potential of OS/2. Mr. Miller observes that, "From a technical standpoint, OS/2 2.0 does appear to be a 'better DOS than DOS.' Of course, the same can be said of DESQview, which is simpler to install, or Windows." This is complete nonsense! Does DESQview or Windows allow one to boot different versions of DOS simultaneously? Virtualize device drivers? Multitask sessions under the control of a preemptive time slice multitasking system? Provide true protection between DOS sessions? Can either DESQview or Windows truly be called a "better DOS than DOS" as PC Magazine concludes OS/2 is? Of course not! The statement that DESQview is "simpler to install" also needs examination. I have known Ph.D electrical engineers who specialize in writing complex computer software who spent days setting up DESQview and QEMM and getting all the device drivers plugged into the correct holes in memory. OS/2 automatically handles all of this tedious setup. The statement about installation is false. Finally, I question the basis for Mr. Miller's conclusion that, "General users probably will not find OS/2 a compelling environment because of the complexity and size of the environment (a full installation requires about 30MB), the relative difficulty of installing and maintaining it, and the lack of many applications native to OS/2." I wonder if Mr. Miller even reads the articles that appear in PC Magazine? OS/2 is inherently designed to be easy-to-use, as explained in the article by Joe Salemi. Even if it takes 30MB of disk space (Mr. Salemi's article states that, "Realistically, OS/2 will eat up about 18MB to 20MB on a average system.") if one wants an easy-to-use, high performance and rock solid system then the price is greater disk space. Many people are willing to get a larger hard disk and some additional RAM if it makes the computer easier to use. And where is the basis for the claim that OS/2 is difficult to maintain? There is no justification given for this statement. In my experience, I find exactly the opposite is true. You just turn it on and it works. And it keeps working. OS/2 is more complex than DOS in the same sense that a Ford Taurus is more complex than a Model T Ford. More complex does not imply more difficult to use and maintain. Mr. Miller suggests Windows as an easy-to-install, easy-to-maintain and easy-to-use system. He states, "Windows 3.1 makes it safe to jump in the water." Question: If Windows is so easy to maintain, then why does PC Magazine run a Windows column covering all kinds of nitty-gritty details of memory managers, PIF files, special config.sys settings to avoid UAE's, and the like? Evidently, it takes alot of fine tuning just to make the system work. OS/2 doesn't require all this constant tweaking - it just works. If Windows is so easy to install, then why is PC Magazine compelled to devote pages 195 to 235 to all kinds of special "tips" to help people get Windows 3.1 working as they undoubtedly expect? Apparently, Windows is no simple matter to get up and running reliably, but Mr. Miller makes no mention of this fact. In the article on OS/2, Joe Salemi states that everything you need to know is a mouse click away on the Workplace Shell desktop. Does OS/2 require special "tips" on avoiding system crashes (UAE's) as PC Magazine feels is necessary for Windows? Obviously not. On the contrary, Joe Salemi reports, "We tried to test OS/2's protection capabilities by deliberately inducing a UAE (Unrecoverable Application Error) in a Windows window. But our attempt failed completely with all of the Windows applications on the test system, including some that will crash when running under Windows 3.0. This is testimony to OS/2's stability." From what is written in PC Magazine, it is obvious that Windows requires a considerable amount of fine tuning to make it work properly. On the other hand, OS/2 is not difficult to install. The statement that OS/2 is difficult to install is manifestly false. Mr. Miller's remarks about lack of OS/2 applications is also questionable when considered in the context of comparing OS/2 to DOS/Windows. OS/2 provides better support for DOS programs than Windows does. OS/2 supports Windows programs. OS/2 supports OS/2 1.x programs, of which there are some very good ones. And there are many genuine 32-bit OS/2 2.0 programs already under development. These multi-threaded 32-bit programs will be far superior to anything showing up in Windows, and they will be available this year. Some major applications, such as Corel Draw!, Lotus 123/G and Lotus Freelance Graphics, are expected within a few months. In the light of these facts, does Mr. Miller conclude that Windows provides more choices for software support than does OS/2? The way I see it, OS/2 offers more choices, and better choices. DECISIONS, DECISIONS Perhaps the most unreasonable part of Michael J. Miller's article is the insert on page 115. First of all, the "Five reasons to buy Windows" should be ADDED to the list of reasons to buy OS/2. Look at what is stated: COMMENTS ON "FIVE REASONS TO BUY WINDOWS" 1. "You can create documents rich with fonts, graphics and more." Same goes for OS/2. Soon you will even able to do it alot faster by using 32-bit applications. 2. " You can run multiple applications." You can do it even better in OS/2 because of the preemptive time sliced multitasking system and Super Fat and High Performance File Systems. 3. "You can combine applications." Same goes for OS/2. 4. "If you do not like Windows, you can always remove it." Again, OS/2's convenient selective install allows you to remove Windows support - the advantage is that you don't have to sacrifice DOS multitasking, improved network support, superior file systems and other OS/2 improvements that are provided to DOS sessions. 5. "All of the major developers are focusing their resources on Windows." Fine. OS/2 will run these programs. Note that many major developers are ALSO focusing resources on OS/2 32-bit applications, so you will have the option to run BETTER versions of their software instead of the old 16-bit Windows stuff. COMMENTS ON "FIVE REASONS TO BUY OS/2" 1. "You bought into IBM and Microsoft's hype the first time around, and now you are stuck." HYPE? STUCK? This is nonsense. Those of us who have ENJOYED two years of genuine high performance multi- tasking (and are not accustomed rebooting our machines several times daily) do not consider ourselves "stuck". We know that what IBM has told us is not "hype". The "hype" was foisted on poor suckers who got "stuck" with Windows 3.0 and suffered through nearly two years of UAE's and deplorable performance. 2. "You need to develop a high-end application that requires 32-bit processing and a real multitasking, multi-threaded environment." WHAT ABOUT THOSE OF US WHO JUST WANT TO BUY AND USE THESE "HIGH-END APPLICATIONS"? In other words, those of us who want to make proper use of the hardware that we ALREADY OWN! Lotus, WordPerfect, Borland, Corel, Micrografx, DeScribe and many other vendors will provide this capability very soon. Don't you think this is worth mentioning? And knock off the Windows NT smoke screen. That is a SERVER operating system. 3. "You need an application that requires multiple communication sessions." What does MULTIPLE mean? I just want to connect to a network, be able to run applications, communicate on a modem and format diskettes without my system crashing, balking or forcing me to become a junior computer scientist. OS/2 has been doing this for over two years. Windows has not, and still does not. 4. "You are an IBM-only shop." Oh, please. This is a veiled attempt to propagate the old myth that OS/2 only runs on IBM hardware. I'm writing this using OS/2 on a Toshiba portable! Why should IBM-only shops have the exclusive rights to a reliable and easy-to-use computer, while everybody else is relegated to use a system based on 16-bit real mode segmented junk left over from the CP/M days. 5. "You are ready to meet the future today." THE FUTURE! Get serious. We have had 386 machines for over FIVE YEARS and are now finally getting an operating system to take advantage of it. This isn't THE FUTURE. This is just getting around to running the hardware we already have at FULL SPEED. Why should I pay a premium to upgrade from a 20 MHz 386 to a 33 MHz 386 when the software is running my machine at HALF SPEED? COMMENTS ON "FIVE REASONS NOT TO BUY OS/2" 1. "There are few applications." Excuse me, but more applications run on OS/2 2.0 than on any other platform in existence. I am sure you are not aware of it, but there are some EXCELLENT OS/2 applications. Try out Freelance Graphics for OS/2. IBM's TCP/IP for OS/2 is as good as it gets on ANY system. PMX (Presentation Manager X-Windows) works great. Mr. Miller further states, "But the reason to choose an operating system access to unique applications." Well, that is ONE reason. Of course, as Windows users will tell you, it doesn't necessarily do you any good if the system can't run these apps EFFICIENTLY and RELIABLY. Also, some of us believe that multi-threaded 32-bit applications qualify as being unique since they (1) Run the 386 at its full potential and, (2) Minimize the appearance of the hourglass. The fact that these will be available IN ADDITION to nearly all the apps that run under DOS and Windows makes it even more unique when compared to, say, Unix, Macintosh, or DOS/Windows. 2. "Unix already offers a 32-bit multitasking operating system." Does Mr. Miller read PC Magazine? Try reading, "32-bit GUI Alternatives: No Contest." on page 193. If you read that article, you will learn that Unix costs over $1000, requires substantially greater hardware requirements, does not run DOS and Windows apps effectively and does not have a single standardized GUI. One could also mention the steep learning curve for a DOS user and the fact that Unix applications cost many times more than DOS, Windows or OS/2 apps - assuming one is available. Unix is not an issue. 3. "You can only run some Windows applications." Well, that is also true of Windows itself! OS/2 will run Windows 2.1 apps. Will Windows 3.0? OS/2 will run nearly all Windows 3.0 apps including (as I understand it) all of the ones on the list of over 30 that Microsoft warns will not run properly under Windows 3.1. Really, this should be listed as a reason NOT TO BUY WINDOWS 3.1. 4. "The installation process is ridiculous." I totally disagree. If you can figure out how to insert a diskette into a floppy drive, then you can install OS/2. In my judgment, it is easier than installing DOS. It does take longer, but that is because the system is doing things that I don't have to do. A great deal of information must be loaded onto the disk in order to make OS/2 easy for a human being to use. This takes time. So what? 5. "You inherit the problems of three operating systems." Actually, you inherit the ADVANTAGES of two operating systems (DOS and OS/2) and three graphical environments (Windows, Presentation Manager and Workplace Shell) while AVOIDING THE DISADVANTAGES associated with all of them. Again, Mr. Miller has made a misleading statement that misrepresents the situation. COMMENTS ON "FIVE REASONS NOT TO BUY WINDOWS" 1. " You have to relearn your applications." Maybe not if you run OS/2! Just run it AS IS and learn more later if you like. This is not always possible with Windows. 2. "Windows is slower than DOS." With OS/2 you have a couple of options. The OS/2 DOS sessions are darned good. Also, when the 32- bit OS/2 versions show up, you can learn them at your leisure and check to see if they are too slow. 32-bit OS/2 apps will narrow the performance gap with character based DOS programs, and should actually be an improvement over graphical DOS apps. 3. "There are still more DOS applications than Windows applications." Of course, as PC Magazine observed, OS/2 is "better DOS than DOS." Also, you can run DOS applications FASTER and MORE RELIABLY under OS/2 than under Windows. So, if this is a reason not to buy Windows, it could be a reason to consider OS/2. 4. "You do not have the hardware." I agree with Mr. Miller's comments here. You get what you pay for. 5. "If you want multitasking, DESQview does a better job with DOS applications." Not better than OS/2 - IF you have enough memory to run OS/2 properly. Again, this is another reason to consider OS/2.