Path: news.mitre.org!blanket.mitre.org!news.tufts.edu!cam-news-feed5.bbnplanet.com!cam-news-hub1.bbnplanet.com!su-news-hub1.bbnplanet.com!news.bbnplanet.com!logbridge.uoregon.edu!newsfeed.direct.ca!news.he.net!vncnews!HSNX.wco.com!waldorf.csc.calpoly.edu!nntp-server.caltech.edu!seqaxp.bio.caltech.edu!MATHOG From: mathog@seqaxp.bio.caltech.edu Newsgroups: comp.os.vms Subject: Beating a dead horse? An exchange with Digital management. Date: 17 Dec 1997 00:39:39 GMT Organization: Biology Division, Caltech, Pasadena CA 91125 Lines: 228 Message-ID: <67770b$4mc@gap.cco.caltech.edu> Reply-To: mathog@seqaxp.bio.caltech.edu NNTP-Posting-Host: seqaxp.bio.caltech.edu After much effort I managed to finally get the PC VMS survey into the hands of the head honchos at Digital, and even had a response from Dick Price, who is "Chief of Staff, OVMS & WNT Systems Business Group". Following my sig are his note (quoted) and my respons to it. Don't bother following up here if your purpose is to support anything that I say, since apparently postal mail is the only thing that gets their attention. (If you disagree, go ahead and post it here, I read this group even if they don't.) Postal mail should be sent to: (person you want Melling, Claflin, Price, Palmer...) (title) Digital Equipment Corporation 111 Powdermill Road Maynard, MA 01754 Just don't use this address for anybody on the Board of Directors other than Palmer, or the letter will come back marked "terminated - remove from mailing list". Regards, David Mathog mathog@seqaxp.bio.caltech.edu Manager, sequence analysis facility, biology division, Caltech ************************************************************************** * Affordable VMS? See: http://seqaxp.bio.caltech.edu/www/pcvms.html * ************************************************************************** >Thank you for your numerous correspondence to Wes Melling, Bruce Claflin >and myself. I will try to address the issues that you've identified. Thank you for your response, but let me be frank (one last time)... > > I reviewed your letter to Wes Melling (same letter sent to Bruce >Claflin) regarding Alphastations and OpenVMS and proposed to you that >those in your survey population that wanted a system for home use, could >acquire a hobbyist license. You wrote back saying you were not >interested, etc. Also, in your response you compare OpenVMS pricing to >that of Solaris and Windows NT saying that our pricing policy and >strategies were leaving our loyal customer base, who are devoted to >OpenVMS in the lurch; and that the pricing was out of line. Those loyal >customers I hope recognize OpenVMS for what it is; the most robust, >secure, available and scalable operating system in the world. That >being the case, it is "premium priced". As with other products, there >are low-end, mid-range and premium varieties. We, and our loyal >customers believe that OpenVMS is the flagship of operating systems and, >therefore, price it accordingly. Being completely blunt about it, you are deluding yourselves about the value of OpenVMS to most of your customers. You can repeat the quality arguments until you are blue in the face, but the sad truth of the matter is that most purchases are first by price, and second by quality, so that quality will win you sales when your prices are competitive, but it won't keep you in the race when your prices are not competitive. Moreover, my survey supports my assertion that your "loyal customers" think OpenVMS is overpriced for the market segment which we are discussing. Monitor comp.os.vms, and you'll draw the same conclusion. Somehow I do not believe that your office is being barraged with calls like "we want to pay a 6X premium over WNT!" OpenVMS is great, but flagship? Let's be honest here. OpenVMS is very rugged, very easy to manage, even inexpensive to manage (if you add in labor costs), secure, and galaxies may make it better for load balancing, but the TCP/IP implementation is rudimentary, the X11 interface is stuck at R5, device support is extremely problematic (for graphics cards, for instance), and software availability is bad and getting worse. It is only for the most sensitive of mission critical operations that the current high prices for OpenVMS can possibly be justified. If I had to create my facility from scratch today we would buy Unix machines from Sun or SGI, even though, all else being equal, I prefer to use OpenVMS systems. That decision isn't based solely on the initial high purchase price, but more so on the lack of native software - these days I have to port nearly everything that we use, other than those few products we get from Digital or from other OpenVMS end users. Years of Digital's boneheaded pricing policies has driven OpenVMS into a niche status, both on college campuses (direct observation) and in industry (judged by the absence of employment ads), and consequently it is becoming increasingly difficult to continue running it. Moreover, the Digital argument seems to be that since it sells large mission critical OpenVMS machines this somehow precludes it from also selling smaller, more cost effective machines, never mind that the smaller machines clearly cannot replace the larger ones for most operations. For instance, who in their right mind would try to run a bank's dataprocessing center on PC class machines? Not enough IO, not enough memory, not enough redundancy. >Additionally, the system that you configured in your survey is a premium >system. You're joking, right? It is a few notches above the bare bones entry level, but it is hardly a premium system. I could buy the Intel version today in any computer store in town, probably for less than the $2176 cited. > I don't know who at Digital Semiconductor told you they could >build that system for that price, the URL is: http://www.digital.com:80/semiconductor/mvi/P010.html In any case, they didn't have to tell us that they could do it, because we know it can be done. The prices are not out of line for the commodity components, with the only real question mark being on the price of the Alpha/Motherboard combination. If Digital cannot design/build the motherboard at at a competitive price there are certainly others who could, and in the aftermath of the Intel settlement, Alpha chip prices should no longer have to include a penalty for the underused FAB. Besides, a few hundred dollars plus or minus on the hardware is not the biggest problem, it is the price of OpenVMS itself that makes it an untenable choice. >but Digital Semiconductor does not >build systems; we in the Digital Products Division do. The pricing is >erroneous (I'm not saying that isn't what they told you, but rather what >the Alphastation pricing algorithm suggests). The 30% gross margin is >also a bad assumption. We spend hundreds of millions of dollars each >year on research and development so that we can provide the most >advanced solutions for customers like yourself and those responding to >your survey. If we were to limit our gross margin to 30% in our >products, the research and development expenditures would be cut >drastically, to avoid a major loss for our company. We are in business >to produce the most advanced computing solutions for our customers and >to attain our profit goal of providing a fair return to our >stockholders. To do both require a gross profit margin of greater than >30%. I think most manufacturers these days would be quite happy to get a 30% margin on the sale of smaller machines built almost entirely from commodity parts. There is just too much price pressure to expect to be able to get more. Companies that insist on higher margins are driven out of the marketplace, as happened to Apple, and is happening to Digital, although you folks seem to be in a state of denial about it. As for those hundreds of millions of dollars of research, well, all I can say is that if it means as a result potential customers cannot afford the resulting machines then it is not money which is being invested wisely. If you must, think of the lower margins as an advertising expense - some of the people who start with those smaller machines will eventually migrate to bigger ones. >Our OpenVMS systems run some of the most critical applications in the >world. They are the standard by which all others are measured. I know darn well what OpenVMS system's capabilities are. but jeez, you guys really don't get it, do you? Not only do most of your potential customers not even consider OpenVMS as a solution, they definitely do not consider it the gold standard. Rather, if they think of OpenVMS at all, they think of it as an expensive, archaic OS for which there is hardly any software - and they are two thirds correct. It's mostly only your existing customer base that still considers OpenVMS options, and for most of us, OpenVMS is becoming increasingly unattractive, not because of OpenVMS itself, but because of Digital and how it prices, and generally mishandles the OS and its customers. >Recently, the organ bank which manages human organ transplants became an >OpenVMS user. They chose OpenVMS because it is the best and in a case >where human lives depend on computer availability, that's what they >wanted. They depend on Digital Equipment Corporation to provide >continuous development and enhancements that will maintain its >leadership in operating systems. To do that we must have the required >funds. You're citing the same market niche as I did above - mission critical, cost is not the primary concern. These sales sound great, but they are just a tiny fraction of the general computing market. Nobody dies or goes broke if my machine is down for a day. Reliability is good selling point, but it is not the only selling point. > >I have found some Alphastations that could meet the needs of some of >your survey respondents that are "value-priced". They are slower >systems than the one you configured, but they are Alphastations that can >run OpenVMS, and do well in a home environment. They are available from >Hughes Data Systems and can be viewed on the Web at >http://www.hughes-hds.com/sales. There are two systems, both priced >under $2,000. They are WNT systems, but an OpenVMS license can be >obtained for them. Hughes Data Systems is a distributor for Digital. Thanks for your efforts, but most of us already know about those. Surely you know that adding the appropriate licenses to make these machines usable would raise their price to nearly $5000? I can put a faster Intel based WNT or Linux machine on my desk for $1700. Even to somebody who really prefers OpenVMS, the OpenVMS solution is not worth 3X the price. > >Your survey was very complete. Thank you for informing us of it's >existence. You're welcome. > Both Wes and I have reviewed all the results and comments. >They have provided us with a unique insight to a very limited number of >our loyal installed base. We aren't that loyal - Digital has not, and is not meeting our needs, and since we apparently cannot convince you of it, we'll end up going elsewhere, as have already most of the (former) OpenVMS users. > There are about 434,000 installations of >OpenVMS today, with approximately 10 million users. I've seen those numbers before, but frankly I don't believe them. I think the former includes machines like the 100 or so VAXes that have been shut down and thrown out on our campus in the last 5 to 7 years (mostly replaced with Suns or PCs), and for the latter "user" is pretty nebulous. For all I know 90% of them are mail users who don't even know that their mail hub runs OpenVMS, and don't have any other kind of account. > Although your >sample was interesting and expressed the views of valued customers, it >was limited. Yes, it was limited, but that doesn't mean that the rest of your customers don't agree, just that they did not see the survey. Do a real marketing study and you will find that most of your customers agree with the "PC VMS" viewpoint - which is why they are buying WNT machines (right price) instead of OpenVMS machines (wrong price). > We know we can do better, and we are always trying. Sadly, I can say with certainty that I am not alone in seeing no evidence of it. Regards, David Mathog mathog@seqaxp.bio.caltech.edu Manager, sequence analysis facility, biology division, Caltech