dmocsny@uceng.UC.EDU (daniel mocsny) (09/09/89)
In article <45369@bbn.COM>, slackey@bbn.com (Stan Lackey) writes: > In article <2108@uceng.UC.EDU> dmocsny@uceng.UC.EDU (daniel mocsny) writes: > >Agreed. But why can't the hardware be tuned to run UNIX? > > I just knew you were going to say that! And, recursively, I just knew you knew, etc., >STACK OVERFLOW< :-) > Actually, it looks to me like > a lot of the newer RISC's are being made for C and UNIX. A trend I heartily applaud. The pace of hardware advance is so fast that vendors lose their market window if they try to go too proprietary with RISC. And with the massive surges in price/performance we're seeing with RISC, watch out for some serious erosion among the proprietary holdouts (but you already knew that...). RISC/C/UNIX could well crush almost everything else out there in the general-purpose computing market, *if* the RISC vendors can coordinate themselves well enough to present a consistent platform to software developers. Price/performance alone might not free locked-in users from proprietary intransigence. The RISC people have to get the applications. The better they cooperate, the sooner they will have them. [ how the computer market got so fragmented ] > If your product is not differentiated in some useful way, the only way > you're going to sell it is through price wars, and you end up with > terrible margins. Competition seems to be analogous to highway speed limits---a great idea, for everybody else. :-) > It would surely be a great thing [for users] if all machines were > compatible, but due to business reasons like these they never will be. OK, let's talk about the philosophy of business. I think the computer industry could benefit from reading what a few others have said about why businesses exist and what they are supposed to do. For example, Henry Ford Sr., in _My Life and Work_ (a must read for any entrepreneur) presented his ideas of how to run a business. Ford claimed that business exists to serve the public, and it does so by continually striving to give the public more value for its dollar. The business that gives the public the most value survives and prospers, and the business that does not, fails. Ford said he did not approach his work with the idea of making money, because that was the wrong way to go about it. He said his aim was to give his buyers maximum value, and then profits would follow as a natural consequence. His book is full of many interesting illustrations of how Ford and his managers continually examined their operations, identifying and eliminating waste, and lowering costs while maintaining or increasing quality. One of Ford's favorite tactics was to lower prices and raise wages at the beginning of the year, and then instruct his managers to obtain the necessary savings. (Too bad Ford's successors forgot all about what he wrote...ah well, that's what happens when an industry matures/stagnates, I guess...) Similarly, Peter F. Drucker, famed management consultant, has written extensively on how successful businesses operate. In essence, Drucker echoes Ford Sr. He says the job of the manager is to (1) determine what constitutes value to the customer, and then (2) to maximize it. What could be simpler? Since you already know where this train is going, I could save the bandwidth and stop right now. But to make sure nobody misses my point, I'm going to spell it out. Computer vendors do not need to hire a bunch of high-priced consultants to figure out what constitutes value to the customer. All they have to do is pull out their earplugs and listen to the screaming. And much of what customers and applications writers are screaming about is the value they are losing because of frivolous product differentiation. I have faith that the quasi-free market will eventually give the customer what the customer wants. If Ford Sr. was correct, somebody should be able to make a lot of money by doing so. I certainly hope that this happens (and it is clearly in the early stages now) before the public wakes up to the fact that horizontal fragmentation in the computer market affects them the same way that any other crime does. Dan Mocsny dmocsny@uceng.uc.edu
gsarff@sarek.UUCP (Gary Sarff) (09/22/89)
In article <2118@uceng.UC.EDU>, dmocsny@uceng.UC.EDU (daniel mocsny) writes: >In article <45369@bbn.COM>, slackey@bbn.com (Stan Lackey) writes: >> In article <2108@uceng.UC.EDU> dmocsny@uceng.UC.EDU (daniel mocsny) writes: >> >Agreed. But why can't the hardware be tuned to run UNIX? >> >> I just knew you were going to say that! > >And, recursively, I just knew you knew, etc., >STACK OVERFLOW< :-) > >> Actually, it looks to me like >> a lot of the newer RISC's are being made for C and UNIX. > >A trend I heartily applaud. The pace of hardware advance is so fast that >vendors lose their market window if they try to go too proprietary with >RISC. And with the massive surges in price/performance we're seeing with >RISC, watch out for some serious erosion among the proprietary holdouts >(but you already knew that...). > I don't applaud it, it seems like stagnation to me. Just as in the PC DOS world, yes dos is "standard", yes it has lots of programs, but also, it is an inferior operating system, piece of hardware. UNIX is not the most perfect, wonderful operating system achievable by man such that we should stop looking for anything new, and yet that is where the industry seems to be headed. UNIX has many flaws, but since it is becoming a "standard" no one dares to fix them, the same old buggy ugly utilities, the same kludgy interfaces to the kernel, poor user administration utils, poor security, etc. year after year. There are "proprietary" OS's that have some good ideas that UNIX users could benefit from, just because something is stamped with the term proprietary doesn't mean it is bad, though that is more and more becoming the perception as the herd stampedes, (or is stampeded) towards UNIX/C. If processors become optimized only for UNIX/C, what happens if someone comes up with some idea, or something they want to research, and it is more difficult/slower/ costlier to test because of this process? Say some new language paradigm that doesn't fit well into the C mold because the risc designers said, "well C doesn't use this so we won't make machine instructions/facilities to do it" Or some new OS design concept. It looks like instead of new and innovative things, we are just going to get faster ways to run incomprehensible awk and sed scripts. This has strayed from the architecture subject, but ultimately the hardware's whole purpose is to run software people's projects. ----------------------------------------------------------------------------- The best swords can cut you and you don't even know you have been cut until you start to bleed.
henry@utzoo.uucp (Henry Spencer) (09/24/89)
In article <00312@sarek.UUCP> gsarff@sarek.UUCP (Gary Sarff) writes: >>> Actually, it looks to me like >>> a lot of the newer RISC's are being made for C and UNIX. >... If processors become optimized only for >UNIX/C, what happens if someone comes up with some idea, or something they >want to research, and it is more difficult/slower/ costlier to test because >of this process? Say some new language paradigm that doesn't fit well into >the C mold ... The sensible RISC designers are building for multiple languages; the C user community is only a small fraction of their potential customer base. Sure, the system is written in C, but all too often the *applications* are written in COBOL or FORTRAN or PL/I (or [barf] Ada). You better believe that outfits like Mips pay attention to this, even if the university researchers don't. Similarly, the sensible designers are a bit cautious about burning anything too hopelessly specific to today's Unix into the hardware. Unix does change. >... It looks like instead of new and innovative things, we are >just going to get faster ways to run incomprehensible awk and sed scripts. >This has strayed from the architecture subject, but ultimately the >hardware's whole purpose is to run software people's projects. Ultimate the hardware's whole purpose is to run *applications*. That means that it is very important to make *today's* applications run fast, if you want to stay in business until tomorrow. The ability to adapt to the new applications that will arrive tomorrow is important, but application speed today is non-negotiable. And applications change very slowly. Even if some radical novelty makes all today's architectures obsolete next week, it will be many years before we stop having to run today's applications. -- "Where is D.D. Harriman now, | Henry Spencer at U of Toronto Zoology when we really *need* him?" | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
jkrueger@daitc.daitc.mil (Jon Krueger) (09/24/89)
gsarff@sarek.UUCP (Gary Sarff) writes: >[UNIX's] same old buggy ugly utilities, See every version since at least PWB. >the same kludgy interfaces to the kernel, See CMU's Mach. >poor user administration utils, See any of the system manager shells. >poor security, etc. See the major commercial operating systems. UNIX is neither more nor less secure than they. -- Jon -- Jonathan Krueger jkrueger@daitc.daitc.mil uunet!dgis!jkrueger Isn't it interesting that the first thing you do with your color bitmapped window system on a network is emulate an ASR33?