mp@mtunf.UUCP (11/25/87)
In a recent article in InfoWorld (Nov.2, 1987), Mr. Stephen Satchell has reviewed a host of 386 Computers, including AT&T 6386 WGS. After a careful study of this article, I could not resist but answer his unfair and totally unjust judgements on 6386. What I write below represent my personal opinions, and in no way related to the views of any other person, company or agency: 1: All 27 machines evaluated and compared in this article are 386 machines. The CPU performance of these machines has been rated satisfactory, good, very good and excellent by comparing these with PC-AT, Model 60, Model 80, and Compaq DeskPro 386 respectively i.e. if a machine performed as good as PC-AT, it is rated as satisfactory, if its performance matched that of Model 60, it was rated good and so on. This rating method is unsound, misleading and unfair. One should not consider a 386 based 16 MHz machine as good if it performs as good as the 286 based 10 MHz Model 60. If that was the case, why will one pay an extra thousand dollars or so for a 386 machine? Note that in this article, AT&T 6386 (that outperforms PC AT by more than three times) is rated as good in the CPU performance. Moreover, by all accounts, IBM Model 80 outperforms Compaq 386 (16 MHz) by about 5% in CPU performance. But even if one goes by this method, a rating of 3.2 is needed to qualify as "very good". ARC 386i with a rating of 3.11, is listed as "very good" whereas AT&T 6386 with a rating of 3.05 is judged to be merely "good". Does the author have any hidden criterion that is known only to him? 2: The criterion that a machine should have at least 2 MB memory on the System board in order to qualify as expandable has no merit. 3: The author could not find any compatibility problems with AT&T 6386 (to quote "Of the machines with the lowest price -performance ratio, this was the only one that had no performance or compatibility problems"), yet he gave AT&T 6386 only a "good" rating for software compatibility. On the other hand, ALR 386/220 that "wouldn't run Framework II due to copy protection problems" (according to the article) received a software compatibility rating of "very good". 4: In the "Setup" criterion, the author states that most of the systems are designed well, following standards set by the IBM PC AT. In starting the review of AT&T 6386 WGS, he writes "6386 WGS is not for the do-it-yourself system builder, but an integrator or VAR could use the 6386 at the heart of a complete package for customers looking for complete solutions". Now there are many issues one can take with these statements. First of all, AT&T 6386 WGS is set up exactly as an AT with menu driven interface, in fact the hardware set up for the 6386 is simpler according to many reports (since the serial, parallel ports, floppy controller are on the system board). Moreover, the author states that the documentation is reasonable for the nontechnical user. Second, if AT is the standard, how can Model 80 with a very different set up interface receive a rating of "excellent". 5: "Warranty": AT&T offers a one year warranty rather than 90 days as stated in the article. In any case, if 90 days warranty is rated poor, a 4 month warranty is nothing to write home about as the author does in reviewing the ACER 1100. To quote "Its low price, super compatibility, and four-month warranty make the ACER 1100 a value leader in this comparison". 6: "Hard Disk Performance" The author writes that the hard disk performance of 6386 is "impressive", yet immediately proceeds to negate the effect by stating that the higher capacity hard disks tend to show up better on tests any way, all other things being equal. One can only wonder on the objectivity of this statement!! 7: Hard disk Parking: The 135 MB hard disk used by 6386 WGS is a high performance, Micropolis disk (Model Number 1355) and is indeed self parking, in direct contradiction to the statement made by the author. In fact all hard disks offered by AT&T on 6386 are self-parking. The disk, in question, is very quiet and does not make a very loud click sound (as AT type disks do) when the machine is turned off. Can this be the cause of the misunderstanding on the part of the author? 8: "I/O Slots" 6386 has 7 expansion slots: two 8 bits, two 16 bits, and three 32 bit slots. The electrical design of the machine allows the XT and AT boards to be plugged into the 32 bit slots. This fact is clearly mentioned in the 6386 documentation. Therefore the article's implication that AT&T 6386 is short of expansion slots is without merit. One can search for and easily find numerous more inaccurate, unfair and unjust comments, statements and opinions presented as technical evaluation, but it is not necessary since one does not need to scan each bit of sand to recognize the beach. The fact is that 6386 WGS is a good machine (not exceptional), and it was unjustly treated by the InfoWorld article. This response is a small step to set the record straight. Thank You all.
wtm@neoucom.UUCP (11/30/87)
I too was unnerved by the Infowold discussion of the AT&T workstation 386. Aside from the out and out mistakes in the article that have benn cited: Infoworld, while giving lip service to Unix, is a blue pin striped msdos slave. They didn't give any considerations to functionality of the machine beyond its ability to run msdos dhrystones (or whatever their bench criterion is). The networking ability and the Unix/DOS merge functions are unique to the workstation 386; they were not considered in the article. These are unique and important assets. Infoworld is too sensitive to price in determining the overall rating of a product. Often a crappy product is given a higher rating than a better fucntioning product, simply because the crappy product is cheaper. People that buy IBM have certainly proved that in business, the cost of capital equipment is not a major issue. --Bill