name: Richard Shockey Date: 1:18 pm Tue Feb 21, 1989

Once again Mr. Blair graces us with his uncommon wisdom and knowledge. Though I am currently not running OS/2 I would agree with his sentiments completely.
Frankly Mr. Canion and company are running scared...it is not self evident that EISA can even work how you combine interrupt drive hardware with what EISA says it can do is ...well I'll believe it when I see it. What you are giong to see is hardware becoming more complex. Read Machrone's column in the latest issue of PC mag. Some precient observations. What is clear is the traffic on the bus is going to het heavy. OS/2 seems to have some problems running multiple asysc tasks effectively watch for the 386 version of OS/2 and its ability to effectively manage bus masters. Some of the wave of the future can already be seen in the new FAX boards with processors and memory built in to buffer the data before the mother processor sees it. Modems will come next in fact i believe the Hayes Micro Channel modem has memory installed. One thing is for sure I would'nt bet against IBM they are setting standards that eventually we will all want. Plus they still make the best keyboards on Earth.

Name: David Blair #13 Date: 10:07 am Wed Feb 22, 1989

has already stated that its EISA machines will be their high-end models. It is also now known that each EISA machine will have proprietary memory cards, unless they decide to get together around the spec of one company, which is doubtful.

EISA was simply a scared announcement made to counter the slowly growing MCA tide that these companies saw. MCA has picked up steam, and, inreality, been stamped "VALID IDEA" by EISA. The BIG difference is MCA is for real, and available NOW, whereas EISA is still a ways off, along with any adaptations to the buss from 3rd party people to make it worthwhile.

Lots of companies have mi-judged IBMs abilities in the past. They still do.

Name: Tom Jorgenson #394 Date: 5:05 pm Wed Feb 22, 1989

According to 'rumors' from company sources, the following EISA 80386 units are supposed to be out this year: Compaq <no date> Zenith Late this quarter Epson This fall Tandy <no date> It's true that they ARE vaporware right now, but they won't be for long. In addition to the above units, I've been in contact with a group of engineers in Silicon Valley who are working on the design for a 50MHz 80386 EISA unit for release late this year to clone manufacturers. These guys have teamed up from design groups who've worked up the boards for a number of existing 80386's. Formerly, they worked on a fixed-cost basis, and the boards were produced under various names. This time they've decided to generate a standard design to be licensed to various manufacturers on a royalty basis (smart guys). EISA really isn't a concern unless you need a FAST unit, such as to replace a minicomputer - and don't mind paying the cost (as with all new technology, these jewels won't be cheap for a while yet). In the meantime, you can get an excellent 20-25MHz 80386 for as low as $3000 or so... Best regards, Tom Jorgenson/First Capitol Computer

Name: David Blair #13 Date: 9:01 am Thu Feb 23, 1989

B: Why do I need to wait until "late this year" for a machine that does no more than an MCA machine has been doing for the last 2 years?

C: If they are building a 50mhz machine, they are using chips out of spec, and I wouldn't touch them. Until Intel announces a 50mhz 80386, that is the only aaumption you can safely make. However, I'll be jealous as hell when they do. I can deal okay with 25mhz machine, and get a tad envious of the 33 mhz ones on the way, but 50mhz makes me green again.

D: I disagree about EISA/MCA being only for FAST machines. I think they are more important if you have BUSY machines- that is, heavily loaded ones, or ones that do exotic work. There is a difference.

Frankly, I still prefer MCA. Less noisy, more sound design, and not a kludge of 1980 technology. Even better would have been NuBus, but you can't have everything...

Name: Tom Jorgenson #394 Date: 7:46 pm Thu Feb 23, 1989

IBM didn't release the MCA buss because it was a better or more sophisticated buss. They released it as a "clone killer". At the time they developed the MCA buss, their intention was to release a new "standard" that manufacturers could only use if they licensed rights from IBM. Thus, IBM could pick and choose its competition, and avoid the beating they received at the hands of the clone manufacturers. In order to make the release of a "new standard" more palatable, they dredged up the premise that the "old buss" couldn't handle "higher speed buss transfers". In point of fact, the original PC buss is quite capable of operating at speeds of 50MHz and above with the mere addition of proper buss termination, and a multi-layer backbone. Also in point of fact, IBM has never actually implemented the higher speed buss transfers that they keep alluding to. Indeed, I would bet from past experience that these features won't actually be supported until the EISA machines appear. IBM isn't actually an innovator, although they certainly have the resources to be one. As a result of the restricted licensing on the MCA product line, IBM has been able to justify higher prices. IBM is smart enough to realize that the base machines have to be competitively priced, but the add-ons carry rather nasty price tags. In addition, innovation on the part of third-party vendors is stifled, since IBM has to bless their products - and they also justify a higher price to pay for the additional overhead involved. Intel isn't currently the only source for processors. Harris is in the process of releasing a number of higher speed processors as a second source - and one of these is a 50MHz 80386. My California friends are not just flakes playing with pushing chips beyond their limits. These are professional engineers with well-known design firms under them. These guys design the Dell (PC Limited) computers as well as a number of other clones (last time I talked to them they were working for Compaq as well). In re: the popularity of the MCA buss, last year an industry-wide poll noted that the most popular of the IBM models was the sole unit IBM had released with the OLD buss. It outsold the other models by 3-to-1. It was immediately after this announcement that the "Gang of Five" released the EISA announcement. They weren't pulling it out of their hats as a result of the popularity of the MCA buss, they were actually dropping it on IBM at precisely the time that IBM was showing the chinks in its MCA decision. The EISA standard isn't a hacked-together standard pulled out of nowhere for the moment. It has been in work for well over a year now. Indeed, EISA stands for "EXTENDED-Industry-Standard-Architecture". The original spec was the ISA standard. As with any industry-wide specification, there are points where all the companies don't agree, but the points are becoming clarified rapidly. Eventually it will doubtlessly become an IEEE standard as well. The impact of the EISA standard can't be ignored. The Wall Street Journal summed up the results immediately after the initial announcement on the front page. As they pointed out, if only the top three vendors in the "Gang of Five" cooperated, their present sales topped IBM's total micro sales by 50%. Within two weeks of the announcement, there were well over 150 companies signed to the EISA specification including Zenith, Epson, Olivetti, Tandy, Compaq, and many others. I've nothing against MCA and I'm certainly not saying that there's anything WRONG with MCA - but I wouldn't buy one myself. Why lock yourself into a single vendor when a standard is evolving industry-wide? Why pay higher prices for the same products? IBM uses FUD to sell products, that's Fear-Uncertainty and Doubt. The MCA issue is being handled by them the same way.

Name: David Blair #13 Date: 9:33 am Fri Feb 24, 1989

A: Thank goodness for OS/2. I had to download that missive and can pop back and forth between sessions to follow it. B: I don't recal noting either you or I working for IBM at the time MCA was proposed or developed, so what their motivation was is, at best, speculation on your and just about everybody else's part. I, for one, don't really care why they did it; it is inconsequential. The facts is that the ISA buss is a noisier buss, and is a much more poorly designed buss. While you can accomplish multiple transfers on the buss, it is not as clean, efficient, or elegant as MCA. And while you may be able to bastardize it to run at "speeds of up to 50mhz" you are diverging from that standard, and no longer are an ISA buss, but a kludge of it. MCA is designed to handle speeds in excess of 50mhz without having to add "proper buss termination and a multi-layer backbone." As to IBM not being an innovator, though they have the resources, I can't agree more, and it is the ultimate frustration I have with IBM. I don't mind their prices (though I wish they were lower), but it really "hurts" to see some of the kinds of things they can do- and for a reasonable cost- go unused or hidden away. As to restricted licensing- I haven't seen much evidence of that. In fact, they have been fairly easy to get a license from. And it hasn't held prices artifically high either- pricing a Compaq puts that to rest. And IBM does NOT have to "bless" 3rd party products. They DO have a center that assigns MC buss ID numbers to prevent 3rd party boards from clashing. There ARE a large (relatively) number of board makers who have simply appropriated numberd from boards that are similar. Those board operate correctly, and all IBM says is "Don't fripe if you get a conflict from your configuration because of non-unique numbering." There is no other restriction. As to Harris making 50mhz 80386s, I am amazed, since Intel has refused to second-source the 80386. I'll believe it when I see it in print elsewhere, from multiple sources. I'm not saying you are lying or wrong, simply that I am skeptical. While last year the ISA buss may have stayed popular, recent tracking shows that the MCA is now selling better for IBM than their old bus models, and that sales are increasing for IBM overall. It was a knee-jerk reaction to IBM's MCA. That's the only way I can see it, inasmuch as it was released prior to its very definition. As to the WSJ commentary ( a fine computer indutry journal...)- there is a massive difference between buying the rights to the spec and using it. Many- in fact most- of those companies have also developed MCA clones. A couple have released them. I don't see buying an MCA machine as locking myself into a single vendor. I have at least 3 choices now, and more coming, for MCA machines. IBM isn't the sole maker of MCA buss additions. Single source logic sin't valid here. Low-number source logic MAY apply, but I can safely say at the moment there are a lot more MCA alternatives than EISA.

Last, while I have a lot of respect for you- I must admit that I have a certain hesitation at some of your perspectives. I have nothing to gain or lose over MCA vs. EISA vs. ISA, while you run a large computer re-selling company which does NOT offer IBM machines. That isn't meant as a put-down, or a comment that you are being less than honest- but merely that your perspective is shaped by your business (as it should be). I'd buy an MCA machine before an EISA. I have nothing against EISA, but simply don't see it as a better solution. Just as a patched program is rarely as good as a properly re-written one, neither is a patched buss. No matter what window dressing Compaq & its band of sheep put on the EISA, it is still 1980 technology.

Name: Richard Shockey #18 Date: 10:49 am Fri Feb 24, 1989

I, for one am willing to take a little wait and see attitude about all of this. We must not forget in all of this debate that people do not buy computers for their busses they buy them to get a job done which usually means software. If the short history of this industry tells us anything is that it is software dirven. Lotus 123 made the IBM pc what it is today the way Visi Calc drove the markets for Apples and Aldus Pagemaker for the Mac. The bus question is only interesting in the context of hardware that does something important, like faster screen writes for graphic intensive applications or managing Async communications. There aresome interesting statistics on Mean Time Between Failures between the old bus and Micro Channel.. IBM (Chet Heath) has allways stated that the main reason for the redesign was not to increase IBM's monopoly power but to vastly increase reliablity and especially cut down on Radio Frequency interferrence. You will notice that all IBM machines are FCC class B while a lot of our clones are At higher clock speeds this is an issue no to be sniffed at lightly. What bothers me is not the EISA debate itself but Rod Canion's (Compaq's CEO) self-rightous attituted "We set the standards..." Microsoft sets the standards by providing people with a common operating system and IBM enforces that with its staggering market power.

Name: Tom Jorgenson #394 Date: 7:12 pm Sat Feb 25, 1989

Dave- While I wasn't working for IBM personally at the time MCA was developed, one of my employees WAS, as an IBM rep. IBM was quoted in a number of publications at the time as stating that MCA was their "clone killer". These weren't my words, they were public knowledge. IBM harrassed us for a couple of weeks about carrying their product line. Indeed, as it WAS IBM I had to give it some thought. To this day I could pick up the line in a minute, if I so wished - I've chosen not to. Although I have the greatest of respect for IBM's capabilities, I don't care for the way they treat the market. The ISA buss is not a dirty buss, nor was it ever. Adding termination to a buss does not make it a kludge - and EISA is not a totally new buss, like MCA, it is an extension of the old buss - and the original cards are still usable. I spent 8 years of my life designing microprocessor based systems in the military-industrial complex, and personally evaluate over 50 new systems a year (including MCA machines) - and I resent your insinuation that I'm pro the EISA buss "because I don't sell MCA systems". That's pure hooey. I promote what I believe in, and what's in the best interests of my customer - even if it means I lose the sale to someone else. Intel was required under their government contracts to provide a second source for their chips. They tried to block these requirement by forcing the potential second-sourcers to abide by a number of contract requirements for sub-licensing - - so that they could say no one was qualified. However, Harris won the sub-licensing under Intel's requirements, and is now producting the lower speed components (and higher speed units in sample quantities). The guys I'm talking about already have their names on a number of very popular clones. One of them worked with Harris on the chip mask design for the 50Mhz processor. It's true that IBM MCA systems have picked up in sales, but not in the same proportion as other units from other manufacturers. In a market that doubles every five years you have to increase sales by 20% each year just to keep up. If you think the major players in the industry aren't really serious about EISA, you are dead wrong. The Air Force is already writing a spec for a multi-user box based upon EISA, and several of the biggest players (including Zenith and Compaq) are already scrambling for a piece of it. I have the greatest respect for IBM, but I don't think the MCA release was in the best interst of the customer. Indeed, over 90% of our business is large business, and most of them perceived IBM's move as an attempt to get them to dump the large quantities of existing PC's for something which wouldn't use the same components - so they shied away from it. Of course, IBM has a number of very loyal customers (which they've largely earned) who will follow them to the ends of the earth, but you have to attribute that to IBM's marketing and support, rather than the MCA buss. In an ideal world, IBM would spend more money on new product releases, and less on fighting the tide - God knows their R&D capabilities are beyond compare elsewhere. It's true that IBM has loosened the reigns on the MCA buss, but they still hold them tightly. IBM made a point of filing a number of patents centered around the buss, and legally a company has to obtain licensing permission from them to release a product which incorporates any of the technology in these patents. In addition, IBM has made it patently clear that they enforce their patents - ref: the lawsuits now going on in Taiwan and the licensing arrangements being forced upon manufacturers there. I really wasn't attacking you, but your view isn't the only one and I felt that it was being presented as a fact where it was really an opinion, just as mine is. My personal feeling is that MCA is interesting, but is a departure from what I consider a PC-compatible - and that has to do with my background as an engineer, not as a salesman...

Name: Tom Jorgenson #394 Date: 7:36 pm Sat Feb 25, 1989

Indeed, you have to consider both software and hardware - and you are indeed right in the Microsoft sets the industry standards. Sometimes Microsoft's power is frightening, what they don't own in software - they control. When Windows become firmly established as part of the OS, it'll get downright scary. I've often wished that Digital Research could have continued to play a more active role in the industry. CP/M was a very crude OS, but their languages were getting pretty darned good. Microsoft is very talented, but I wish they weren't so willing to use ever-more-RAM the way they are. What I think is happening in the industry is that IBM is heading back to selling micros the way minis have been sold - with a hardware-independent interface, and that the rest of the industry is heading towards a real vendor-independent standard - EISA. I think this might well splinter IBM from the rest of the world for better or worse. If Windows comes into its own accord, this may not matter much because, application software will run on both types of machine and continue to do so. IBM is, and always has been, great at getting their customers to perceive fault where there often is none. Let's suppose you purchased an el cheapo PC clone, and it proves less than reliable. Then IBM announces that MCA is designed to remove the "design flaws" of the old buss. Geez, musta been the buss, right? Let's remember that IBM designed the PC buss in the first place, and gave us the AT buss as well. At the time, the busses were clean and reliable, right? The PS/2 machines are very interesting in many respects. I especially like the way IBM was able to make the "snap-together" modules in them, and the manufacturing capability that went into them. The 3" floppies are also beginning to grow on me as well. They certainly have their points. However, I think it would have been much nicer if they hadn't created a new buss standard, and had pursued an extension of the current buss standard instead. This would have been more in line with their stated purpose of continued product compatibility. MCA uses a multilayer backplane for noise reduction, and will have to use buss termination as well if they decide to push for higher buss transfer speeds. These are standard techniques for extending a buss' operating speed - and will have to be used in any really fast machine. For my part, I'd rather not toss out the many upgrade cards I use in my computers just because IBM wants me to go a new way. I'd also prefer not to pay 50% for new cards when I buy them for the same reason, just because of IBM's patents. (shoulda said 50% more)

Name: Tom Jorgenson #394 Date: 7:40 pm Sat Feb 25, 1989

In re: Rod Canion's comment that "We set the standards...", I wish he were right. We sure can't count on ANSI or ISO or IEEE to set them - - at least not until the market has already made it's choice... |-)

Name: David Blair #13 Date: 9:45 am Sun Feb 26, 1989

B: To a "t", every person I know of in IBM and associated with IBM says that the buss was not created as a simple marketing ploy, but, rather, they had plans for their line thatt could not effectively be dealt with with the older, noisier buss. That was their own words, not mine. They said that the noise factor on the buss made it such that the shielding had to be continualy added to keep it clean from the exterior; interioir noise was also a problem to them. I have seen the same comments echoed in technical journal after technical journal. As to IBM's way of business, I can't argue with you there. As to your taking offense that I feel your perspective may bot be totally biad-free, I'm sorry you feel insulted because it was no more meant as an affront to your personally or professionally than anything you have commented back to me. I have previously- in fact as recently as last night- recommended you & your business to friends & associates as an excellent source of machinery. If you take offense, you do so where none was intended. I still, however, cannot help but feel that you business interests have an effect on your perspective, conscious or not. Its the same way I feel about Rod Canion. I can hardly expect him to stand up and say "Yes- MCA is better, and we plan to convert!" Give me a break. You want a marketing-decision driven buss, look at the EISA. Its vaporware of the worst kind- panic attact announcement of a standard they'll develop as soon as they do. As to the additions to the ISA buss- I still believe them to be kludges. They are additions to an admittedly inferior buss that are being grafted on. All the comment about it being otherwise won't convince me. I also find that the Air Force has any specs requiring the new buss to be nothing more than humorous AT THIS TIME. Anyone who makes a requirement something that isn't a reality yet is asking for trouble. Maybe they'll require 1-2-3 Version 3.0 also. Start a new category of workstations called "vaporstations." Yes- I realize both EISA and 1-2-3 3.0 will be here eventually. But they aren't here now. As to Harris picking up as a second source, I haven't seen it and thus can't say anything about it. I don't doubt you; I just am unaware of it. Can't comment on it. Funny- while all the people like to point out little things about IBM sales, I feel VERY safe in saying MCA is vastly outselling EISA, and will have a larger installed base for quite sometime. You can do that when the product is for real, and available, as opposed to some nebulous "Late this year" time-frame.

I have no doubt IBM did it to increase their sales without the competition; that is simply good business sense, and not uncommon. Funny that when other companies do the same thing, they aren't shot down for it. I keep forgetting that the rules of business sense don't always apply to IBM the way they do to everyone else.

And OHMYGAWD! IBM is enforcing their patents! Heaven forbid they go after what is theirs! I can't believe they'd do something so vile and mean. After all- isn't it IBM's responsibility to do R&D for all these other little companies so they can get a good jump into the race against IBM? To find something wrong with that REALLY makes me worry. If IBM has a patent, they have the legal and moral rights to recieve proceeds from them. They aren't a charity. They're a business. Granted, a damned cold, humorless one, but still a business.

Volley to you.....

Name: David Blair #13 Date: 9:53 am Sun Feb 26, 1989

Have to agree with you as to some ramifications of the MS power. I am worried that no one will be ther to counter-check them in what they do. While I am a proponent of OS/2, I see LOTS of little niggling things they could have done better or more smoothly. Ditto with a lot of their other products. Like many companies, they have started doing all their work in C, and developing it on pseudo-machines from which they port to the real machine. Makes for portable code, but also makes for highly bloated, inefficient programs.

EISA vendor independent? Surely you jest? EISA will be as much under Compaq's control as they can make it. They'll have rights to everything developed for it. And as to IBM designing the older buss- yes, back in 1980/1983 (AT). Betcha engineering has made advances since then. Betcha ISA hasn't. It was clean at 8mhz, 10 mhz. Started getting noisy past that.

As to tossing out cards when you upgrade- I sold my machine and all but one of its cards when I upgraded from my IBM AT to my Everex (can't say I'm defending MCA because I bought one!). At work, all the new machines we are ordering (surprise! I am hoping to be able to direct where they do their buying when budgets get approved, and one plae I would consider shopping is your place...) will have all new cards, peripherals, etc. And using those old cards gets you none of the benefits of the new buss.

And what about those proprietary memory cards????

Name: David Blair #13 Date: 9:54 am Sun Feb 26, 1989

Thanks, but I'd rather have someone else besides Rod Canion set standards. I don't know who yet, but not that guy.... I'm as leery of him as I am of Mr. Pepsi and the fine IBM people. (Now there;s a statement that should confuse you- that I am leery of IBM people....)

Name: Richard Shockey #18 Date: 10:17 am Mon Feb 27, 1989

Maybe we could call ESPN and sell the rights off for this. Now for a little independent commentary while our contestants are in their respective corners. For those of you new to this.. Tom...certainly no one is impuning your business judgements you carry the Epson line and as a user of those machines for the past 3 years with NO repeat NO problems I am anxious to see how they implement the EISA system. But I still agree with Mr. Blair in the sense that this business was founded on software compatibility not bus structures. For all of IBM admitted faults their sheer market power gave us the environment we live in. It may be be the best there is...but we could be stuck with the problems in the UNIX would if there wasnt some agreement on standards dictated by market dominance. I repeat we really could have 6 or 7 bus standards so long as we still had a single operating system ...well two (DOS OS/2) could have anyone other than IBM/Microsoft created a new operating system that business would accept for their PC's. NO of course not what this industry need now is not a secondary debate on what kind of bus is the best ...but some genius who can put 8megs of memory in a box for $200.00.

Name: Tom Jorgenson #394 Date: 11:18 am Mon Feb 27, 1989

The Air Force has already SEEN EISA machines. You see, companies like Zenith will provide evaluation pre-production units for qualification testing prior to actually running them in production. These are not sloppy wire-wrap models, but have all the features of production units. Zenith has a box called the Z-1000 which is an EISA machine now being shown to large buyers of micros. The Air Force was quoted as being "very impressed" with the box - although the rest of us will have to wait to see it for a few months yet. IBM public comments versus their private ones are quite different from each other. They will "leak" comments to the press regarding what their real intent is. They will then find a "justification" for the "real reason" why they are doing what they are. They will then carry that concept through as a "policy". That is, even their employees will be told that "this is the company line" and what should be said publicly. Thus, the justification becomes reality. Why do they go through such a round-about way of getting information out? Because it gets them press time. They've found through long experience that it pays to set up "trustworthy" members of the press to talk about "what IBM is REALLY doing". I have several writers for national publications that are personal friends - you'd be surprised how much they know that they can't release until a specific date. Magazines like BYTE have 4 month lead-times for publication. Haven't you wondered how they managed to keep the magazine current? IBM has been designing "high-speed busses" for years now. They didn't design the original PC buss (which isn't the ISA buss, by the way) to be a "slow buss" or a "noisy buss". The only things they left out that prevented the raw buss design from working at higher speeds was a multilayer design (to put power and ground in the card center for shielding) and buss termination. Incorporating these features is a snap and very inexpensive, and the resulting buss is quite comfortable at mugh higher buss transfer speeds. Naturally, the MCA buss incorporates these features - but that's no big deal, especially since they aren't actually USING the higher-speed buss transfers as yet. Re: the issue of what we sell. It isn't that "we promote what we sell", it's that "we pick up the products that we believe in". If we believed in MCA, we'd pick it up in a minute. Re: IBM's patents, I have no objection to IBM protecting their patent rights My point was that in order for a third-party company to produce a product incorporating MCA technology, they have to obtain licensing from IBM. What that means is that IBM has control over "who sells what", and that the smaller start-ups can't afford even the basic research it takes to see whether the licensing is required or not. Thus, IBM has limited the competition to a trickle. If you don't believe this to be true, compare the number of add-on MCA cards to those for "the old buss", and... while your at it, compare the prices as well. IBM is perfectly within their rights, but they still have put a damper on the development nonetheless. Maybe an interesting thing to do, once the EISA's are released, would be to bring an EISA and an MCA into a meeting, and take them each apart. I'd be happy to supply the EISA machine for the purpose. Would also be happy to run benchmarks, although it might be an unfair comparison, since the EISA's being newer will doubtless have higher speed clocks that existing MCA models. Best regards, Tom Jorgenson/First Capitol Computer

Name: Richard Harnack #19 Date: 10:41 pm Mon Feb 27, 1989

information and points of views. Just one or two things from the non-technical end for a moment. 1. Regardless if EISA or MCA is "the way to go", IBM and others will have to begin to market tyo people like me more effectively if they expect their "standard" to become "standard". While I do want to get the best machine I can for my money, I also want to know that service and warranty is available for a resonable length of time after purchase.

1. As far as I can see, OS/2 software really will not make a big impact for another 3 -5 years, by which time the market will have sorted out which piece of technology is "the best".

2. Technical aspects aside, will these machines be comprehensible to those of us who are still running 8088's? This is where marketing becomes more crucial. IBM is definitely not going out of business in the next year or two, how many clone makers are still around now that were here two years ago? This is going to be the main challenge for those embracing EISA, will they be here next year?

If it were only as simple as deciding on the technology. -Richard-

Name: Dave Kelewae #398 Date: 12:05 am Tue Feb 28, 1989

As a common buyer fo computers I feel that price and compatibility are going t o drive the market. Most people don't ever need a 50mhz 386 to run most if not all of what they want to do. And don't you get tired of shelling ouy money every year for a new computer? Give us a break.

Name: David Blair #13 Date: 9:17 am Tue Feb 28, 1989

B: I'm not sure how Zenith has a model for show of EISA since the standard is just now being finalized. I would assume it is a "should be like this-may be a bit different" type of prototype. As t your comment on IBM business practices, we haven't disagreed, so I don't know why you return to it. As I said, the people I know from IBM all told me this on a personal level- this wasn't media reading. The old buss was put together with a 1980 mindset- the EISA is a 1980 mindset with some 1988 ideas glued into it. I've seen high-speed implementations of the buss and it just gets noisy as hell. It echoes terribly (sort of like what radio transmitters do with SWR for ham oriented people), drops interrupts, and bogs down heavily. It underutilizes DMA. Most of these things can be corrected, but you have gotten away from the buss, and it becomes a kludge. No matter WHAT window dressing you put on it, anything as heavily patched as EISA is, boils down to being a kludge. MCA is simply a better method. You aren't restricted to running the old cards at the old speed, and you can accomplish more with it. There are some nice additions that EISA has that are slight improvements over MCA, and it is nice to be able to save you old net cards, modems, etc. I just don't think in the loing run it will pan out. The ISA and EISA busses are tied completely to Intel architecture. MCA is not. That in itself is cause to think. As to running EISA against MCA machines to benchmark- it depends on if you look at Tandy, ALR or IBM machines. Second- the fact that EISA will be "newer" (a debatable idea in and of itself) doesn't mean it will be faster- we have no idea what MCA machines will be out in between.

Frankly- I wish that they all had settled on a true 32 bit standard- NuBus. A better idea than both of them. And I don't mean the screwed up bothcing that Apple did with the Mac II series- but true NuBus. Unfotunately- in both cases, NIH principles ("Not Invented Here") applied...