It has only been two weeks since the first HP e3000 Community Meet, but I already miss the mates I met there. We were all serious about being in that hotel meeting space on a Saturday morning, a time when soccer games and family activities could call lots of people away. Keynote speaker Jeff Vance had to duck out after lunch to coach a game. Donna Garverick and her husband James Hofmeister had similar Saturday duty, but the well-known couple of 3000 advocates showed up the Friday night before for supper.

Many people might wonder what we were all doing in one place, talking about a computer HP canceled so many years ago, with a serious share of its community already migrated or on its way. We gathered to be thinking together, learning and sharing ways to steer the future forward, toward our desires.

See, through that sharing we could hope to transform that ledge the community feels — even now, more than six years after the worst November news most community members would hear — to turn that ledge into an edge, and see what might be, to grab the advantage of imagination.

Not long ago IBM announced its purchase of Cognos, the creator of the development language PowerHouse. The language continues to drive HP 3000 applications in places too diverse to track down, but there's a lot of PowerHouse still running out there.

David Ivey of id Enterprises says that the IBM purchase is a clear signal that the 3000 customers who use PowerHouse need to get serious about moving off their 3000s. The acquisition is both "an opportunity and a spark" for his company which has specialized in HP PowerHouse development through the past two decades, as well as its more current offerings of migration, Ivey says. Primary targets for migrating PowerHouse sites are Windows solutions, especially employing Visual Basic and Visual Studio.

"IBM's purchase once again emphasizes that the 3000 has a limited lifetime, and you need to make preparations to move off it," he said. "You need to have a plan and get busy."

[Editor's note: XML is an important data exchange technology that can use a focused tool suite for HP 3000 customers. The entry that follows explains how the latest XML solution gets the job done in the MPE/iX enterprise. Making a business case for using XML is detailed on our Nov. 27 entry.]

By Peter Prager

XML, which is a self-documenting, flexible data format, should give organizations an edge by providing reusable and repeatable standards which can be utilized throughout their business processes. XML Thunder, an XML suite from CanAm Software, visually enables all data mapping, conversion and error handling required for reliable XML-processing. This frees up developers to focus on business issues — a  time-saving advantage when compared to compiler-dependent solutions.

XML has many benefits, but it is a complex technology featuring hundreds of rules. XML program development is an ideal target for automation as the rules are numerous, but very well defined. An XML handling program must implement all relevant aspects of the W3C rules with stable, efficient code. This code has to provide appropriate handling of simple and complex types, XML structures, features and error processing. More importantly, it has to be robust and reliable in order to handle mission critical data that organizations rely on for their business.

How does it work?

XML Thunder generates COBOL and/or C sub-programs that can be invoked from a main business program. These sub-programs are referred to as XML Handlers. XML Handlers do the actual XML processing on data received from the main program. The mainline interfaces with the XML processing sub-programs through a special data area called the Interface Data Structure or IDS.

There are two types of XML Handlers, XML Readers and XML Writers. An XML Reader is used when the content of an XML document needs to be made available to and processed using traditional COBOL or C data structures. When using an XML Reader the mainline may obtain the XML document in any manner which is suitable, e.g. MQ, RDBMS or a sequential file. Once the XML document is held in memory, the mainline will pass it to the XML Reader subprogram that was generated using XML Thunder. The XML Reader will parse the content of the XML document according to the design time rules and will pass the content back to the mainline in regular data structures for further processing.

An XML Writer is called from the main business program. Then an application needs to create an XML document. The main program will pass data using regular data structures to the XML Writer, which will construct the XML document from the data based on design time rules specified.

The generated code is entirely ANSI COBOL/C source code, without any hidden runtimes or executables.

Think of XML Thunder’s flow as consisting of three major steps: Select, when files are selected containing the XML schema/DTD and/or COBOL/C structures; Bind, when COBOL/C variables are mapped to the desired XML nodes; and Generate, when COBOL or C program code is generated implementing the XML Reader/Writer design.

By Peter Prager

Data exchange between different platforms, such as HP 3000s and Windows workstations, is complex, requiring significant knowledge and investment.

As vendors have tried to further protect their intellectual property with proprietary data formats, the challenge of exchanging data in a heterogeneous computing environment involving a variety of platforms has been growing over the years.

This issue has been substantially heightened with the B2B use of Internet technologies, where a large number of different platforms need to share data. This is especially true when considering the increased requirement for national language support.

XML stands for eXtensible Markup Language and was defined by the World Wide Web consortium in mid-1996, based on the SGML meta language. As XML’s main purpose is to enable any data in any national language to be shared across different computing platforms, it has increasingly been a natural solution of choice.

The fall months brought far more than a change in weather and the colors of leaves. This season has served up many a transition for your community and the computer world as a whole, but none was so profound as the 3000’s Extra Two Years.

I was glad to have this online vehicle to keep up with the news all through October and into November. Not a bit of irony escaped me as I noted HP’s announcement date for a critical data corruption patch: Oct. 31, the day celebrated as a Worldwide Wake for the 3000 four years ago, the last day HP sold the system.

So now 48 months have passed, a full year of them beyond HP’s first “we’re getting out” date, and the community remains on the system in large numbers. About a third are already departed from your community, their migrations or replacements or re-hosting complete. But the rest of the user base needed those critical patches, those that HP announced primarily over the Web on newsgroups, Web sites and mailing lists. HP has sent word by way of the postal system, too. Quaint but comprehensive.

OpenMPE is not some government entity with a roster of voters to look up and scrutinize. I saw a membership roster in 2003. Only 125 names back then, but not even half were vendors. Consultants, some. But also customers.

As for the voter names, I get a list of those each spring to oversee the annual election. I cannot share those, but yes, even in their limited numbers, they represent the viewpoint of a customer trying to get a workable solution for the near term.

No tax dollars are not at work at OpenMPE, any more than they are inside HP. Member and voter identity is protected on a privacy basis, just like customer lists at HP and third parties. Insisting that this group of volunteers mimic the government’s information revelation is not an unusual request. But it’s just not going to happen, until these volunteers get support contract monies from 3000 sites. That will make OpenMPE accountable to their customers, in some way. That may never happen, given these HP extensions to its support business. On the whole, the extensions look sensible until you consider their impact on OpenMPE’s desires.

As for what the organization is doing right now, several times a month the board meets by phone conference with HP’s Jeff Bandle, as they did with Mike Paivinen before him. Projects such as OpenMPE review of the HP MPE/iX build process investigation — “hey, what does it take to create a new MPE/iX build, or a patch?” — that’s the kind of project OpenMPE is involved in. One such project has already taken place. OpenMPE wants the 3000 group in HP to be involved in two more such reviews.

Also, OpenMPE got HP to announce the extension of support earlier in 2007 than HP had planned. You may recall that the 2005 announcement — where 2006 didn’t mean 2006 anymore, and the companies who hurried and spent plenty learned they should have waited and spent more gradually — that 2005 announcement got dropped on the community and customers during the holiday hiatus. Bad publicity, awful exposure.

This year’s announcement came much earlier, during the period when 3000 sites were planning 2008 budgets. The OpenMPE calls are all covered under executive sessions. HP insists. HP holds all the cards in this relationship.

The flaw in approaching the 3000 market from a spot of selling mostly migration solutions is that the solutions and viewpoint becomes a hammer, the kind that makes every 3000 problem look like a nail. “Just migrate,” people say, “and it will all get safer and easier.” But the migration experts who are ready to help, they know — a migration is the most complex IT project a shop will undertake. Some take a matter of weeks or days. Others take years. They are fraught with pitfalls and learning. Most of them make the Y2K work seem simple by comparison.

Homesteaders have limited budgets, operate systems often frozen down, don’t test patches, take few upgrades. This has always been a belt-plus-suspenders marketplace. Homesteading is the easy path of no change. Shifting a 15-year-old application suite to new code, new compiler, new database, new OS, that can feel like cowboy stuff to a 40-plus 3000 director.

And there’s no hurry or firm deadline, especially if HP support has become a thing of the past for a customer. By my estimate no more than 1 in 3 HP 3000 customers now take MPE/iX support from HP. This is a marketplace accustomed to a decade or more of value from an enterprise server purchase. The current generation of HP 3000s, the A- and N-Class servers, are only finishing their sixth year of active service in the market.

So what’s Open about OpenMPE? There’s an open dialogue with HP about how to make the homesteading, whether interim or for an indefinite future, easier and more affordable.

Maybe OpenMPE will have no future with HP, in the long term. But nobody else in the community is taking up the prospect of how to create patches during 2009 and 2010, when HP quits. And to think that nobody will need a patch for a 3000 in that period is a bit innocent, if you ask me.

So who is OpenMPE? The directors, almost exclusively, people whose pictures and names and bios are up on the Web site. As well as those who have served on that board in the past, for free. If OpenMPE went away tomorrow, how would the 3000 community, which will be working at least through 2010, benefit? The group’s executive conference calls help HP plan the details of its end-game. HP says so, and says the group has impact. And after the disgraceful melt-down of Interex, the 3000 community needs some kind of HP advocacy.

On this day in the US, we give thanks for what we love in our lives with a big meal, connecting with family and friends. It's a rare thing for Abby and I to be away from our Texas home and NewsWire's offices on this day. We're usually hosting holiday guests, but this year we're on the road, visiting my family in a little town on the leftmost edge of Lake Erie.

As I sit here in an Extended Stay Suites room, waiting for all the feasting to come, I am reminded of that E in Erie. The lake's name still draws sniggering, maybe the same kind you've heard about the HP 3000 in your shop. Yes, it is that system older than all of the rest, but as vital as a Great Lake with freshwater. Something common but essential, and yes, something to be thankful for.

That E also reminds me of something Alan Yeo said about the HP 3000. HP started to call the system e3000 back in the year 2000, a marketing move to prompt a new look at legacy technology. Yeo wrapped up last weekend's 3000 Community Meet by saying that e stands for enduring. It's one of the many e-things we can give thanks for as 3000 community members:

• Extensions, of support by HP to keep the vendor officially in the 3000 business, and of support from the third party suppliers and 3000 gear from resellers. About the only thing really missing now is new systems.
• Exploration, by the OpenMPE board, still seeking a way to extend the problem-fixing patch process once HP leaves the community
• Excellence, from the 3000 solution suppliers who have built products with durability and rugged design, like Adager's Alfredo Rego says, as if they are deep space satellites which must operate for years without need for maintenance, so your IT duties are more manageable
• Enthusiasm, from the vendors discovering new ways to migrate several decades of business logic to new environments. Everybody tells us that the application makes the most difference in choosing a new platform. The platform makes the most difference, however, when the application is already written and reliable on the 3000 — and it requires patience and innovation to carry it into the future.
• Exactitude, from those caring for a 3000, either by proxy or at your site, as well as the exacting development of solutions to mirror the talents of IMAGE, MPE/iX on other hardware, and more. Excellence is required.
  

As for Abby and I, we give thanks for the Excitement of covering a vendor's End-game, to chronicle the Evolution of your community to an independent Entity. Enjoy the holiday.

Another two years, HP says, and it will be closing up its HP 3000 support business. Will that 2010 date mean the end of migrations, or homesteading, or the community? Not by my figuring.

I like to do some fun figuring here when I glance at the calendar. By December 1, The 3000 NewsWire will have pursued 3000 news and delivered issues and blogged reports longer in the post-cancellation period than before it. And 2008 promises to be a fun year to keep telling stories. After all, HP’s still going to be in the business, supporting us and you.

After all, why give up now? HP’s 3000 support business is a profit center. Until its resources retire or shift to other HP projects, keeping those doors open in a limited way is an easy choice — or easy if the decision-maker works in the HP Services group.

But this makes the OpenMPE initiative worry, concerned that yet another two years will elapse before HP must keep its promise and turn over the keys to the MPE/iX kingdom. That move would give OpenMPE a chance to become more than advocates. Right now, some community members see OpenMPE and homesteaders as worn-out horses, thoroughbred but as blinder-ed as any aging racehorse relegated to harness racing.

This is a short view, by my analysis. First, OpenMPE isn’t any less open than HP’s 3000 business group, which cannot say how many engineers still work there, what number of customers continue to receive PowerPatch tapes, how much business the group books, and a host of other details. Millions of dollars are sent to HP on behalf of 3000 customers, but the business is no more open than HP decides it should be. It’s a profit-driven operation, just as it should be.

But oh my, put OpenMPE in the same harness and its critics revile the volunteer effort. How can it be open when the member list is unknown? What are the group’s true objectives? Is this group just a handful of vendors clinging to the past, hoping to continue to make money off MPE?

In order, the answers appear to be
1. It can as open as it wants to be, just like HP and its 3000 group.
2. Its objectives are the same as they were nearly six years ago — to make a life for the 3000 beyond HP’s business schedule.

Please note, HP’s extensions of its 3000 support now total almost as much extra time as HP first allotted to the system’s lifespan (five years promised, now four more years extended.) Or as Gavin Scott put it at last weekend's e3000 Community Meet, the extension period now represents 20 percent of the entire HP 3000 lifespan.

HP on Nov. 19 posted results from its latest quarter which broke its single-quarter revenue numbers by nearly 30 percent. HP promoted the push as led by the software gains in the company's operations, according to CEO Mark Hurd. A $28 billion finish in the final 90 days of sales for the the year gave HP a total of $105 billion for its annual sales. Q4 netted $2.6 billion in profits. Year-long profits for 2007 were $7.2 billion.

“Strong performance across our businesses was highlighted by sharp improvement in our software segment,” said Hurd. Overall, he noted, “We have added over $12 billion of new revenue this year. While we still have more work to do, HP is well positioned to make further progress in the marketplace.”

In comparison, Apple's last four quarters netted $3.5 billion in profit on one fourth of HP's revenues. IBM has posted $12 billion in profit over its last four quarters on $118 billion in revenues.

HP also announced that its board approved the repurchase of $8 billion in its common stock. This would represent about six percent of all outstanding HP shares. HP started the trading day at a price of $50.87 per share, about three dollars off its 52-week high.

Highlights from the report, released after the markets closed, showed that the Enterprise Systems and Storage Group, (ESS), the parent group where the HP 3000 alternative systems such as Integrity and NonStop servers are created and sold, posted a 10 percent revenue growth over last year's Q4.

The Business Critical Systems group, heartland of the Integrity alternative, increased revenues by 5 percent in the final period of fiscal 2007. Integrity sales grew 59 percent in the quarter, but HP lost ground in the PA-RISC and Alpha sectors, places where the vendor has already promised an end to the server lines built upon those chipsets.

BCS profits rose over last year's Q4 to $693 million, 13.5 percent of revenues. But industry-standard server sales dominated the ESS increases for the quarter. Overall these Intel-based systems sold 14 percent more than in last year's Q4, sparked by x86 (Xeon) blade revenue increases of 78 percent.

However, BCS numbers paled when compared to the HP Services operations, where the vast majority of HP 3000 customers still do business with Hewlett-Packard. HP Services grew to $16.4 billion on the year, an uptick from the 2006 results of  $15,6 billion. Services covers nearly every HP product line, but the BCS yearly revenues fell for 2007, down from $3.6 billion to $3.5 billion.

They came on a soccer day, kids's sports day, a morning for extra sleep. The Saturday of the first e3000 Community Meet dawned with a filled room of the  who’s who of the 3000 community. Or, as QSS application founder Duane Percox joked, “who’s left.”

Who was left had lots to share and report. At left, Rick Gilligan (standing) of bank app vendor CASE shared his MicroFocus COBOL experiences with Bob Cavanagh of Acucorp, now part of MicroFocus

The meeting reached its 50 attendee target, a larger share of partners and consultants than customers. But for each seat of the 50 which were filled at the Doubletree Hotel on San Francisco’s bayshore, a few, a dozen or a hundred other HP 3000 customers were represented. The nine hours of networking, counting breakfast, lunch and the after-meeting supper, offered a view of the future with several destinations — and no certain end of life.

HP was well represented at the meeting, a full table of engineers, managers and experts sitting on the front row. Jennie Hou, HP’s e3000 business manager, updated the crowd on the vendor’s new offerings during 2007, a talk she delivered to far fewer 3000 community members during the summer's HP Technology Forum. Of course, back then, HP's limit to its 3000 business was the end of 2008, not 2010.

The brightest sparkplug of the meeting, ScreenJet's Alan Yeo, (second from left, along with Transact creator David Dummer, Eloquence's Michael Marxmeier and Cavanagh) concluded the meeting with an insight. "I've often wondered why HP placed the e in front of 3000," he said. "Now I think I know. It stands for enduring, because the 3000 has lasted a darn sight longer than HP expected."

HP's contribution to the day was hard to overlook. For all of its IO expert firepower, community liaison connection and even Ross McDonald's lab director oversight, the most moving expression of the day from HP came through a former employee who kicked off the meeting. Jeff Vance had lived in the HP mountaintop, and literally worked from there in his most lauded and productive years developing from home, before retiring in May.

Just six months after his HP exit Vance stood before us to testify about how alive the future feels for him, all while a collage of 3000 division team members flashed on the screen behind Vance. The photos celebrated the retired and departed while making a case for how the end of life of work cannot be calculated when you love what you do.

Lunch delivered the biggest torrent of networking, the brightest benefit which attendees could carry away from a surrended Saturday. A room buzzed with revivals and reports. Some of the brightest lights in the community exchanged updates of the future outside of the 3000 extension, as well as the personal reports on migration tricks that make networking such a solid extension of online relationships.

   HP 3000 advocates at OpenMPE believe the transfer of source code to the volunteer group should begin now. But HP’s 3000 group officials continue to keep their calendar deadlines intact for MPE/iX licensing. HP intends to license, but scheduling is another matter, HP officials report.

   Jennie Hou, the business manager for HP’s e3000 operations, said that the open letter from OpenMPE’s board has been read and discussed by HP’s 3000 management team. However, HP will not be following any proposed November announcement of a timetable to license MPE/iX source in 2009.

    That process is not being accelerated even though HP will halt patch development for the HP 3000 during 2009. The availability of Hewlett Packard support for the 3000 customer, in any form, will determine the start date for any licensing of MPE/iX to third parties. HP recently extended its support for MPE/iX through 2010.

   “Other than altering the timeline to 2010, our intent [to license] stays the same,” she said.

   HP announced in December 2005 that it intends to license parts of MPE/iX to interested third parties when basic support is not available from HP any longer. But just because HP calls its 2009-10 support “Mature Product Support without Sustaining Engineering,” rather than Basic Support, doesn’t change the timeline. The transfer clock still starts when HP exits all of its 3000 support business, according to HP 3000 community liaison Craig Fairchild.

   “The intent has always been that when HP is no longer in the MPE/iX or HP 3000 support business,” he said. “That’s when we would proceed with this licensing endeavor.”

By Birket Foster

    It’s not just about the 3000. Like a facilities manager, an IT manager has many kinds of budgets to consider. A new strategy, one that treats applications like they are corporate assets, can deliver valuable benefits to help a company using HP 3000s move forward.

   IT people have been notoriously bad at getting their place at the boardroom table. An IT budget needs to follow the business direction. The Application Portfolio Management (APM) strategy ensures a good business fit for all of a company’s programs.

   Studies show that managers spend 80 percent of the IT budget maintaining their current assets. If you are forced to do anything radical you run into real issues, then overrun your budget. At most companies, the IT budget is set at operating level.

    Migration can be a radical step. But the duty of an IT manager who oversees a 3000 is to keep track of what is productive. It’s not about the migration, it’s about the whole portfolio. You must assess the 3000’s risk versus the rest of the applications in the portfolio.

    At M. B. Foster, our team has been doing migrations for the past five years. We actually started with data migrations in 1985 — but once HP announced its phasing out of the HP 3000, we began working on the methodology for migrating applications as well.

   Our typical discovery process is to look at a company’s business, both today and where it is going. We then evaluate the viability of each application in supporting the business going forward. Many customers have applications on different platforms and in different stages of supportability and business fit.

    What do we mean by business fit? It is how well the application supports the needs of the business. We ask each department using an application to explain how they use the system and what is missing. Is there a formal wish list for application changes? In a typical situation there may be two years of backlog needed on the application.

Only once in the history of The 3000 NewsWire has a specific date appeared on our pages. The day rattled through your HP 3000 community six years ago on this very day, when HP announced its exit from three decades of the HP 3000 business.

Four years ago, HP 3000 customers and friends around the globe held a World Wide Wake for the system, gathering to raise a glass in toasts and revive the memories around more than 30 years of success using this computer.

Three days from today, 50 or so of the community's most curious and connected members will network in San Francisco over a weekend. The fellow sparking the e3000 Community Meet, 2007? That would be ScreenJet's Alan Yeo, who also inaugurated the idea of a wake for your system on that 2003 day when HP stopped selling the 3000. (There's Yeo helping at a conference, showing a way to migrate COBOL apps.)

On this day in 2001, the badge of "homesteader" was born on our pages. We had to call the majority of the community something, and "non-migrator" just didn't feel right. Neither did the decision to cut off a good product line that wasn't growing as fast as the HP CEO wanted. But we've all moved on from that day, haven't we? You're learning what's next, or gathering your independent resources to homestead awhile — through 2010 and beyond, by all reckoning.

Just today out on the 3000 Internet newsgroup, a hardy soul offered a bit of gallows humor about returning to the 3000 community. "Welcome to the bread lines," he said, as a 3000 veteran announced his return to the newsgroup's membership.

We don't want to get too religious here. But if the life of the 3000 customer in 2007 takes place in a breadline, it might be one forming behind the "loaves and fishes for all" kind of line. The ecosystem looked shocked to get its early obituary in 2001 from HP. Today it looks like a long line of companies ready to help you go or stay. This was my fervent hope in that dark week of 2001 — that you all would rally and keep your own counsel about the right time to move along to new horizons. It's much harder to break up a community than to take a product off a price list.

When HP advised you to migrate from your 3000 — on that fateful day we all mark tomorrow — it might have had something like the latest Windows-Intel hardware system in mind.

There's a new line of quad-core servers based on the Clovertown version of the Xeon 5300 chipsets. If you recall, Xeon is that other processor that Intel manufactures. It outsells the Itanium, well-regarded by HP and the 3000 customers who've migrated to Integrity server, by a wider margin than anybody can calculate.

But popularity is a poor measure for anything except market longevity. (You don't need to tell HP 3000 owners about that.) The new Xeon systems will be 50 percent faster than their dual-core predecessors in what HP calls its ProLiant line. (Extra processors never add 100 percent more performance per addition, another fact that's not new to a 3000 community that operates many 2-way and 4-way servers. Quite a few of them drive mission critical applications, but that's not the point today.)

HP will ship nine of the quad-core models in all, and like much of the ProLiant line, the hardware acquisition cost is often lower than buying used HP 3000 systems. There's also the cost of managing a Windows environment to consider when making a migration. But lots of 3000 sites are already paying their Windows resource surcharge, albeit on the desktops of their workers.

In this case, cost of ownership includes the factor of popularity. People who gather comfort from building up their enterprise on industry-standard platforms choose Windows to replace HP 3000s. They do this in your community more often than they select HP-UX and Integrity, which is why the newest round of ProLiants is significant.

Reports across the Web today say IBM will tender an offer to purchase Cognos, the maker of PowerHouse development software for the HP 3000 as well as Windows-based Axiant, the development environment Cognos offers to migrating HP 3000 sites.

The deal will put $5 billion on the table to acquire Cognos, a company whose product line and IT focus has settled squarely in the business intelligence marketplace for many years now. SAP and Oracle recently snapped up BI providers, and the Cognos stock moved from the middle $40s to above $50 a share since Oracle's acquisition of Hyperion Solutions.

IBM offers to pay $58 per share in cash for Ottawa-based Cognos, a 9 percent premium over Cognos' $52.98 closing price on Friday. The stock settled in at a closing price of $57.15 after today's trading. In an IBM press release, the company said that

Following completion of the acquisition, IBM intends to integrate Cognos as a group within IBM's Information Management Software division, focused on Business Intelligence and Performance Management. IBM also will appoint current Cognos President and CEO, Rob Ashe, to lead the group, reporting directly to General Manager, Ambuj Goyal.

With HP announcing its exit from the 3000 market, Cognos recently offered a mature product support plan through 2009 for the MPE/iX version of PowerHouse. HP followed suit about six weeks later with its own Mature Product Support deal for the 3000 and MPE/iX, through 2010.

Reaction from the PowerHouse mailing list about the acquisition has been limited to customers and long-time consultants as of this afternoon. But one Canadian user of PowerHouse tipped his hat to an historic partner in the 3000's legacy. Cognos, after all, began as Quasar Systems in the middle 1970s, selling the first independent report writer for the 3000 as its only product. Quiz made Cognos possible.

"Mind you, quite a success story," said Guy Werry, Senior Systems Analyst of Hudson Bay Mining & Smelting Co. Ltd. On the mailing list he posted, "From a smallish consulting company with an in-house report writer to a $5 billion company in about 30 years? Not too shabby."

HP enterprise user group Encompass announced the results of its annual election, a vote that resulted in just one 3000-experienced director on the new staff of volunteers. This is not necessarily a setback for the 3000 community, since former Encompass directors stay linked closely with the current board.

Chris Koppe of Speedware is not on the board any longer as a result of the election. Koppe was instrumental in keeping the 3000 community ties to a user group alive during 2005. Interex went bust in that year, and Koppe was on that group's board at the time, and he and then-president Kristi Browder found ways to attract Interex members — many of whom were left as unpaid creditors in the Interex bankruptcy.

As an example, the attendees at this weekend's e3000 Community Meet could purchase a 1-year Encompass membership for just $35.

Steve Davidek, former Interex advocacy committee member and a manager of the HP 3000 shop at the City of Sparks, Nevada, remains on the Encompass board. Davidek helps to oversee the community survey which Encompass uses as a measure of its productivity.

OpenVMS got all the breaks back in 2000. The operating system has been around since 1977, just a few years younger than the HP 3000. But at the turn of this century Compaq decided to bring VMS into the Open realm, re-engineering the OS for the upcoming and then-much-derided HP Itanium architecture.

It was so early in the Itanium game that the servers running the architecture were several laps behind Alpha processors (on the Digital side, owned by Compaq) as well as the PA-RISC systems (powering the HP 3000s and HP-UX servers). None of that mattered to the Digitial brain trust. They could see the advantage in getting their customers a new architecture for VMS to run on. That's what you do for an OS that at the time was 23 years old.

MPE was nearing its 30th year, if you count the aborted release of 1972 which ran for just hours at a time. HP, which created Itanium along with Intel, considered making MPE work on the newest computer architecture. But at the time, the VMS community had something that MPE's customers did not: About 10 times as many sites.

Critical mass, the tipping point for computer vendor business decisions, did not spark HP to proceed with the tough engineering to get a new vessel for an OS nearing age 30. Digital did make the leap, than found itself acquired by HP within a year or so. (The irony was rich, if you were around long enough to recall the wars between DEC and HP in the 1980s. Hewlett-Packard was struggling to get out of the 16-bit processor family with something called RISC. But it was still a few years away. DEC crowed, "Digital Has It Now." Fifteen years later, Digital had it again, while HP didn't give Itanium to MPE.)

Now, an article in Information Week celebrates the 30th anniversary of the VMS operating system, adding that "Hardly anyone noticed."

Thousands of VMS fanatics noticed, but a group much smaller than the Unix devotees or the Windows warriors. Perhaps hundreds in the MPE and HP 3000 community took note, if only because that 30-year celebration, as the former SIG-Sysman chair Scott Hirsh noted, "coulda been the HP 3000." At the least, because the 3000 has 20 more years left in it, until the CALENDAR intrinsic rolls over in 2027 and quits working.

Okay, we admit it. We went off on a bit of a rant today in our podcast, letting off steam about the stream of sniggering at OpenMPE. A few sniggers, maybe, but deserving of some response. After all, who else is taking care of the relationship between HP and the 3000 customers who will remain, relying on the system, once HP exits this community?

Maybe another vendor in the community, one that hasn't raised its hear. But for now, OpenMPE is the best you've got. Listen to our 15 minutes on the subject (its download time probably takes a fraction of the download as the latest Vista Service Pack). This is still a good market, for the homesteaders using the system as well as the migration experts who want to help a good share of the community exit. Going into 2008 with an incomplete migration, well, that qualifies as a homesteader.

End games need to have plays drawn up, and this game has been going on for more than 30 years. The 3000 deserves an end at HP as admirable as its success for the vendor which created it.

HP made it very clear last week: Binary patches created for the HP 3000 are a common Hewlett-Packard repair for a 3000 bug. HP doesn't consider a binary patch any less reliable or more poorly engineered.

How interesting, then, to hear last week the comment from Allegro Consultants' Stan Sieler that his company, a Resource 3000 partner, could have produced a binary patch to fix the Large Files corruption problem. Sieler took note that Allegro could do so, but held back to wait on HP to produce a patch that could be tested more completely — and perhaps integrated more closely with the rest of the patches HP is developing for the MPE/iX environment.

HP had to reach down into the millicode that drives the move_fast_64 call in MPE/iX to repair the problem, for those of you who want to know every detail about the 3000's internals. The HP fix was better, Sieler noted, because it could be plugged into the rest of MPE/iX more easily. By HP, for now, and until the end of 2008. What happens beyond that date — even if HP creates more binary patches like this one — is a matter for OpenMPE to negotiate with HP.

With 12 days left until the e3000 Community Meet, Bayside 2007, only a dozen spots are open for the free lunch and priceless networking opportunities in San Francisco. The community members who have signed on are among the most expert and senior. Solutions to improve homesteading reliability, or avoid unnecessary costs and delays in migrations, will be on hand Nov. 17, a Saturday.

Registration is open at www.hpmigrations.com/sfevent. In addition to the modest $20 collected for a commemorative shirt, the attendee can get a budget $35 one-year membership rate to Encompass, the only remaining user group for HP enterprise customers.

Migrating customers will find the most use from the Encompass membership, especially if they plan to attend next year's HP Technology Forum. Frankly, nothing of this scope has ever been assembled in so short a time. It was only August, scarcely three months ago, that the Meet was being dreamed up by ScreenJet's Alan Yeo and Marxmeier AG's Michael Marxmeier.

HP responded to our Nov. 1 report with some corrections and clarifications about its new patches to avoid data corruption while sorting Large Files. In summary, the patches in question have passed beta testing and are now general released (GR), the sign that their quality satisfies HP's requirements.

What's more, HP recommends that customers install both the millicode repair patch as well as the patch to address issues with MPE/iX SORT processes. Craig Fairchild of HP wrote me this evening.

HP's strong advice is for customers to install both patches. You are correct in asserting a high priority for MPENX11, since it is the patch that addresses the issues with SORT and the MPE/iX OS. However, MILNX10 is also important to address the possibility of continuing to use the millicode in question. Even if a customer is not using Large Files today, there is no guarantee that they won't experience growth that will cause their files to cross into the large range at some later time.

What my initial report got wrong, according to Fairchild, was calling a binary patch testing status into question. Binary repairs like these cannot "skip steps," as I wrote in the Nov. 1 blog headline. If anything, HP says these critical patches got extra testing.

The first is the blog title, “Binary bug patches can skip steps”. This simply isn’t true.
There is a repeating of this theme throughout the article:

  “...but the nature of the problem... might require more than binary patches.”

“these fixes will present a challenge to application developers who will need to integrate them into MPE/iX at some point.”

“HP didn’t do the usual testing for these repairs, it appears.”

“Binary patches like yesterday’s don’t have to cross the user testing and MPE/iX source code integration hurdles.”

While we call down the creative thunder of the promised podcast for this week, we decided to update the process to apply the latest HP patches, software the vendor has labelled critical for all HP 3000 customers.

OpenMPE's director Donna Garverick-Hofmeister clarified the need for reboots to apply those Large File fixing patches HP released on Oct. 31. One patch won't need a reboot, she says, while the other needs a reboot but can be staged.

Oh, and the no-reboot patch, to avoid data corruption in very big files, cannot be staged, according to Donna.

The HP 3000 millicode patch (MILNX10A), the first in 16 years, is needed to repair access to any in-house applications that have used Large Files, or do a sort with a temporary file that can exceed 4GB. If your app has not been modified since March 30, 2000, it's safe. That's when HP introduced the Large File feature.

Large Files has been engineering which HP has been trying to remove from customers' 3000s. A November 2006 patch was designed to turn off Large Files and get those files on the system converted to Jumbo files, much better engineered.

Article revised Nov. 2, based on HP responses

Yesterday's report on a critical patch for HP 3000s sparked immediate response from the user community, especially those experts in the 3000's internals. These experts believe the nature of the problem might require more than binary repair patches, to serve the long term needs of 3000 customers.

[HP disagrees, and its response is available in a Nov. 2 report]

HP intends to create these binary patches through 2010, as it said yesterday. OpenMPE advocates say they are concerned that these fixes will present a challenge to application developers who will need to integrate them into MPE/iX in the future. OpenMPE wants to do this work.

"We've done as much testing as we could get done," said HP's community liaison Craig Fairchild. "There has been some field testing, and a lot of in-house testing." He added that HP scanned its internal 3000 applications to test the abilities of FILECHEK, which finds Large Files.

Our report of yesterday may have been too cheery about the chance of hitting this corruption bug. Our Oct. 31 story estimated that in one case, users risked just an 800 million to 1 chance of hitting one of five bytes at the end of a sort of a 4-billion-byte file that could corrupt data. Stan Sieler at Allegro Consultants, a Resource 3000 partner, said it's not that rare.

We found it fairly easily. 133 file sizes from 2 to 32,766 bytes per record can directly encounter the problem ... but only for Large Files. (As it happens, 256 bytes isn’t one of them, nor is any power of 2.) We stumbled over one while doing a large sort.

Files of about 2 GB or more, and of any record size, can encounter the problem while being sorted, because HPSORT creates a scratch file whose record size isn’t identical to that of the input file — if it happens to create a file with one of those 133 file sizes, and that scratch file is 4GB or more, then it can run into the problem.

Most people don’t have Large Files ... they could be affected only if they happen to sort files that are bigger than about 2 to 3 GBs.  (And, even then, there are only a about 152 file sizes out of 32,766 that might trigger the problem.)

Systems without Large Files are safe. Adager's Alfredo Rego posted to the HP 3000 newsgroup about an hour after HP announced the critical patches. Rego said his lab found the problem in August, after an earlier HPSORT patch introduced a new problem while not solving the initial bug.

Rego's exploration also raised the question of a more serious issue in the nature of HP's repairs, one echoed by MPE/iX veteran and MPE-education.com co-founder Paul Edwards.

"There's some strong issues here," Edwards said. "I'm concerned that these binary patches were not done in the source code. So HP didn't generate a General Release patch, which means it may or may not be tested on all three of the supported versions of MPE/iX. There was no beta patch, or anything."

[HP said in its Nov. 2 reply that the patches have both been beta-tested and General Released.]