Number 72 (Update of Volume 7, Issue 5)

HP user group endorses merger without polling members

Acting independently of its users' preferences, the board of directors of the HP user group Interex announced it will be voting the organization's HP stock shares in favor of the merger with Compaq.

Interex sent its members a notice which reported the seven board members voted unanimously. The organization's thousands of members were not polled by the board before its vote. A prepared statement released by the board said that "We believe our members will enjoy the inherent benefits of a marriage between HP and Compaq," according to board chairman Bob Combs. "The merger of these two industry pioneers will give Interex members broader access to leading technology, R&D programs, marketing, and service and support, enabling them to strengthen their computing platforms."

Online criticism of the Interex board's move appeared immediately. Alfredo Rego of database utility provider Adager, a member of the group since the 1970s, said on the 3000-L newsgroup that "I do not like it when people claim to speak for me. The Interex Board may have voted unanimously, but such is not the case with Interex's membership. If I need to express an opinion (or a "belief"), I do so myself. Likewise, I do not like to endorse anybody else's opinions."

Some members called for an immediate poll from the user group following its announcement. A majority of its volunteers are associated with the HP 3000 community, a market which HP is moving away from by the end of 2006. An early Interex poll indicated a majority of its HP 3000-using members didn't support HP's decision to drop the 3000 platform. The organization is still conducting an online poll on the merger from its Web site only, but results have not been released to support the board's announcement.

The Interex board's announcement was part of a March 11 story on the Web site of financial newspaper Financial Times. FT.com said that Interex is "the largest group representing HP customers." But some customers said the Interex board's show of support doesn't represent their beliefs.

"The announcement implies that the board is speaking for the Interex membership, which is not the case," System administrator John Clogg of Coldwater Creek said over the Internet. "At the very least, the board should announce that its decision in no way reflects the opinions of its members, and that Interex is not endorsing the merger."

We're against it -- merger will do HP's enterprise no good

Holding absolutely zero shares in either HP or Compaq, we've used our 18 years of analysis of HP's enterprise computing strategy to advise against the HP merger. The recent decisions of the HP 3000 division to walk away from the platform show us that products which don't follow in lockstep with industry trends have a dim future at HP. We don't believe there is much hope for sustaining distinguishing technology choices by following trends so rigorously. Commodity computing is the goal of the merger -- a strategy which deprives customers of unique, efficient product designs simply because the products don't follow industry trends.

Using the company's recent HP 3000 decision as a reference point, sometime in the future other non-standard HP technologies will diverge from industry trends. The merger hastens this day, combining technologies and products from HP and Compaq which compete, most seriously in the enterprise computing sectors of the two firms. Product cancellations will follow in the wake of the merger, shortening the lifespan of HP platforms which the company and some partners are advising HP 3000 customers to take.

Building a larger computing company won't deliver HP any significant competitive advantage against rival IBM. If that were the case, the US Federal Trade Commission and the European Commission would have looked more closely into the merger. Neither regulatory group blocked the deal, and the Europeans didn't even ask for additional information. It's clear that bigger won't mean better.

HP 3000-owning companies, as well as managers employed by them, control HP shares. Like HP, we believe that every share matters in the merger. We believe it's in customers' best interest to vote no on the merger, to save HP the resources the company might spend on a merger -- and deploy those more resources in supporting 3000 customers through the Transition. While we don't believe blocking the merger will reverse HP's decision to depart the 3000 space, voting no sends a message to HP to mind its current enterprise business during a critical period for 3000 owners.

We were wrong: merger likely to linger beyond March 19

Despite our fondest hopes in the most recent FlashPaper, it looks like March 19 won't be the last day for debate about the merits or folly of the HP-Compaq merger. Although that's the deadline for proxies to be received, counting could well descend into a Florida-like pace, given the closeness of the vote and the thousands of proxies to be examined. In a Wall Street Journal article the firm which will do the counting, the four-person IVS Associates, reports it expects to look at votes from 900,000 shareholders. The count could be conducted by hand in case of disputes, a distinct possibility considering dissident director Walter Hewlett is spending $34 million to campaign against the deal, while HP is spending even more.

The proxies will travel on March 20 from HP's shareholder meeting in Cupertino to the IVS offices in Delaware, where the counting will begin. The IVS quartet must determine if each proxy is the shareholder's only vote, and if it's the last vote. Shareholders can vote multiple times; the latest dated proxy is the only one that counts.

And there's every reason to believe the vote will be very close. After the ISS advisory to its customers last week recommending to vote for the merger, one of its 23 institutional shareholders voted contrary to the advice. The California Public Employees' Retirement Systems (Calpers) cast its 7.6 million shares against the merger, another .39 percent of the outstanding shares. Calpers said it fears the merger will cause HP to lose focus on its core strengths.

The day before Calpers revealed its vote, Standard and Poors lowered the rating on HP's senior debt. S&P cited HP's pro forma risk profile in making the move, a profile that assumes the merger with Compaq will go through. With a week to go before the shareholder meeting, the 25 institutional investors tracked by the Wall Street Journal were voting shares right at the 2:1-in-favor margin HP needs. The company must have 61 percent of the remaining shares voted yes to offset the 20 percent of shares already blocking the deal from HP's founding families.

But plenty of shares remain undecided among the largest investors. Still unpledged with a week to go were the 2.41 percent of shares held by State Farm, one of the largest users of HP 3000 systems in the world. That's a company that recently learned HP won't be supporting the company's investment in MPE technology beyond 2006.

Hurry to vote on the 3000 improvement ballot

The 3000 community started what may well be its last chance to impact HP's enhancement of the platform, as user group Interex posted the System Improvement Ballot online at the group's Web site . The ballot contains 21 requests that anyone can vote upon, even if the customer isn't a member of Interex. The first listed improvement request asks the HP 3000 division to facilitate "a way for users and developers to continue to run MPE in some supported fashion after end of HP sales and support."

The ballot is being conducted only online and voting ends on March 24. Customers can vote up to 20 points total for their most-needed enhancements. However, the Interex voting software was preventing customers from casting all 20 votes for the first listed enhancement request. A 10-vote limit was being imposed for only that item on the ballot. SIGIMAGE/SQL leader Ken Sletten convinced Interex to deploy a 10-vote limit on the first item, to force voters to hold back some votes for other items on the ballot. SIG IMAGE/SQL has nine items on the 21-item ballot. Another ballot item, number 19, asks HP to "Port MPE and its subsystems to Intel or create an MPE emulator for Linux. In the alternative, enable a third party to do so."

7.0 and 7.5 -- is there time to get them right?

Even if HP gets specifics on making the 3000 a better system this spring, some experts believe there's not enough time left to make any such enhancements bullet-proof and debugged before the division's October 2003 end of life. HP architect and OpenMPE board member Jeff Vance has said that as far as he can see, the division only has until that date to continue to work on the operating system for the 3000.

Mike Hornsby of Beechglen, the independent support company now taking on 3000 sites moving away from HP support, said even using MPE/iX 7.0 could pose risks given the time left in the division's lifecycle. The problem, according to Hornsby, is that too few customers are putting 7.0 into production to uncover bugs. The forthcoming MPE/iX 7.5 will have an even smaller customer base, since the release's major enticement is expected to be native support of Fiber Channel peripherals.

HP may not have enough customers running these two releases to create patches, Hornsby said. "Patches are more a function of customers running a release in critical mass than anything else," he said during our February issue Q&A. "Even if new features made it out on the 7.5 release today, there's not enough time between now and the end of 2003 for enough people to install it for a critical mass testing of the code, in my opinion. We take a much more conservative view of what's a production release."

If Hornsby's advice to his company's supported customers holds, then HP may be spending the rest of its 19 months it will ship 3000s sending out systems which don't meet such standards for production. That viewpoint may have a serious impact on how many sites choose to upgrade to A-Class and N-Class systems, since those computers require MPE/iX 7.0 or later. The conservative approach to production grade systems might also keep customers from signing back up for HP support to receive 7.0 or 7.5. HP still intends to collect back support from 3000 sites which want to rejoin HP's support system and get the latest release. Those back support charges might be the first HP has tried to collect for a system which it already has announced it will walk away from.

Migration tracks swell at Symposium

Interex was keeping open its early bird rate for the e3000 Solutions Symposium later than ever this month, with the lower-cost rates expiring on Thursday, March 14. The training event April 2-6 is chock-full of advice on migration from the platform, as well as a smaller set of seminars on sticking with the 3000 and exploring the options offered by OpenMPE advocates.

As a Platinum Plus sponsor of the event, Lund Performance Solutions will conduct a special a half-day interactive session on its migration services from 8 AM to noon on Thursday, April 4. Lund promised to show "how the experienced e3000 and Unix experts at Lund do system and code migrations. Lund's system specialists, in conjunction with their European partner Open Seas, will describe in detail how they perform the pre-migration systems assessment, develop a plan and solid budget proposal for the most efficient and cost-effective migration path, execute a smooth transition to the new environment, determine the final hardware configuration, and complete the process with a post-migration performance analysis. Included in the presentation will be several detailed case studies. Lund is also offering two additional hours of training in Migration Planning at the conference.

Options for customers Homesteading on the HP 3000, or waiting to follow their application supplier's Transition plans, are also a part of the show. Lund is offering an hour of Ecometry Database Optimization and Performance Tuning as part of the event's new Ecometry Track. QSS's Duane Percox will teach about using COBOL in Web applications, Taurus Software's Vicky Shoemaker is giving a Data Warehouse 101 primer, and HP's Mark Bixby will lead courses on using Perl, sendmail, Posix and the Secure Apache Web Server. That last item was promised as a free enhancement to the customer base by HP during the latest HP World conference -- perhaps a reason to take on the 7.5 MPE/iX release, should Secure Apache emerge there.

Users who registered online for the conference by March 14 got a $695 rate, probably less than four nights of Bay Area accomodations is likely to cost most attendees. Later registrations cost $100 more. Interex was advising customers who register online to use Priority Code SS02PE4. The Symposium promises to be the most 3000-focused event of the year, and we hope to see all our readers who are attending at the show. 3000 NewsWire issues will be distributed to all attendees.

Does LDAP give HP-UX a chance at IMAGE-like speed?

Several issues ago the NewsWire's intrepid experimenter Curtis Larsen outlined the advances a 3000 user could enjoy employing the system's LDAP capability. (Okay, it was in the August issue last summer) A new Larsen exploration, this time on Python, is in our March issue.

Larsen updated us recently with news he heard about HP exploiting the LDAP advantage to get its non-3000 systems to perform as fast as an IMAGE database. The Knowledge DataBase (KDB) was the large one HP once had in its Atlanta Response Center for support and call documentation of HP 3000 systems. Larsen reports:

"The guy I talked to said that HP came up with their own LDAP schema for the data, then internally published it running on HP-UX servers. Since the DB data is mostly read-only anyway, using an LDAP directory on it and distributing various pieces of the tree to different systems would work quite well. It was mentioned tongue-in-cheek to me because this was the solution HP had to come up with in order to beat (or maintain) the performance level of the old IMAGE database-on-a-private-volume solution -- because none of the relational DBs they tried could approach the performance they wanted within the cost constraints (well, duh!)"

Larsen added, "This was particularly interesting to me because of my love and excitement for LDAP, as it presents a practically perfect example application for how to use it (and migrate an existing application to it). Man, I'd shine shoes to get a hold of that schema!"

Of course, if speed is what you'd like to get a hold of, you could always stick with the IMAGE/SQL original, instead of using the LDAP imitation on another platform.