HP 3000 customers were impressed by the scope of Hewlett-Packard's roadmap
for the system presented in late January. Those involved with development
for the systems
found plenty to hope for in the HP message, but some saw the promises as a
sign of
declining impact for MPE/iX projects at HP.
HP's strategy gave specifics into 1999, and some viewers saw a full five-year plan offered. Any promises beyond the current year indicated a strong future for the platform, regardless of their term. John Lee of Vaske Computer Solutions, an HP 3000 reseller based in Minnesota, said HP sounded more commited to the 3000 than he's ever heard any company sound about NT or Unix.
"HP committed to the 3000 for five more years -- in this industry, isn't that equivalent to about a lifetime?" Lee asked. "Has Microsoft publicly committed to NT for the next five years? Has anyone committed to Unix for the next five years? It seems that we're quick to criticize HP for not committing longer, but we forget that no business in their right frame of mind commits to anything far in the future."
Specific commitments in some areas were sparse, particularly in the compiler enhancement arena. Cecile Chi, an independent consultant and analyst based in Pittsburgh, said she didn't hear enough about what HP will do to improve Transact or even the more widely used COBOL II.
"I did not hear support for current ANSI Cobol standards, I heard HP say 'we're talking to SIGCOBOL', " Chi said. "HP knows perfectly well what SIGCOBOL wants, but is not committing to anything. I heard support for Transact, but I did not hear 'enhancing Transact to support the announced B-trees in IMAGE or pop-up windows in VPlus.' "
HP's plans for COBOL need to include COBOL 97 enhancements, she added. "Since just about everybody uses COBOL, I was really hoping for something like, 'HP will begin work on implementing the new COBOL standard when it is finalized.' They didn't have to commit to particular features or a time frame on a standard that's not out yet."
Jim Phillips, MIS manager for Therm-O-Link, said he wanted HP's commitments in the broadcast to include support for both COBOL 97 standards and the extensions that make COBOL II a well-integrated tool for HP 3000s.
"We want to see the HP COBOL compiler support the ANSI 97 standard when it is promulgated," Phillips said. "At the same time, HP needs to continue to support the HP extensions that are vital to making maximum use of COBOL, particularly intrinsic calls. This is not just idle curiosity for us. We are betting our business on the HP 3000 and COBOL."
Some promised enhancements had a dramatic impact for customers. Glenn Cole of Software al dente in Cupertino, Calif., was impressed with the demonstration of connectivity to NT networks through Samba. "The quick Samba demo was a high point of the teleconference for me," Cole said. "In 15 seconds, I finally understood what all the fuss is about."
HP's decision to nudge MPE/iX toward 64-bit support with a move to use files greater than 4Gb seemed sound to Michael Gueterman, president of the consulting firm Easy Does It Technologies, Richland, Wash.
"I can see the business logic in their current approach," he said. "Outside of greater than 4Gb file sizes, the only other thing I can think of as a real benefit is the increase in a flat memory address space, which would help Allbase and Oracle."
Some buts on 64 bits
Customers and channel partners had mixed views on the lack of a
commitment to
IA-64 chips for future HP 3000s. But most customers were willing to accept
HP's promise
to take another look at Merced support in the future if business needs
demand it.
Richard Gambrell, Associate Director of Xavier University's Information Technology Center, said HP's view of its futures were broader than he's seen in the past. His site uses both HP-UX and MPE/iX systems, which gave him a perspective to compare both the 3000's past commitments as well as HP 9000 plans.
"Compared to what was said last year and the year before, HP has gone out on a limb to commit to continuing enhancements and upgrades for MPE/iX," Gambrell said. "If they wanted to close the door, they could have been much more tight lipped. They announced a much longer range strategy than anything you can find about HP-UX"
Like many customers, Gambrell heard CSY general manager Harry Sterling fail to close the door on Merced support for MPE/iX. "Harry didn't rule out a 64-bit version [of MPE/iX] nor a migration to the Intel/HP processor," he noted, "just that it didn't appear to be necessary or useful to meet the HP 3000 customers' needs."
"If the HP 3000 community can gather the strength to keep the business use of the HP 3000 strong and growing, of course HP will port to 64 or 128 bits. The PA-8xxx chips are 64 bit chips and they're not bad. Who needs the Merced chip for business transaction processing? The Merced isn't even a proven winner."
Other customers and channel partners weighed in on the side of confidence in the 3000's processor and 64-bit choices. Jeff Kell, curator of the HP 3000 Internet mailing list and one of the architects of the Proposition 3000 movement, said continued improvements in the PA-8000 line are a good barometer of the 3000's health.
"As long as PA-RISC continues to evolve we're safe," said Kell, who's the HP 3000 resource manager at the University of Tennessee at Chattanooga. "I would hope to see a mirror-image porting effort between PA-RISC and IA-64 for both HP-UX and MPE. If that remains absent through the end of this decade, when announced PA-RISC committments currently stop, I'm concerned"
Bob Lewandowski, vice president of systems for ASAP Software Express of Buffalo Grove, Ill., said that support for 64 bits in his HP 3000 operations "isn't really an issue for me. I've got growth potential with what they've announced. My concern is to give me better development tools. The Java Development Kit is big news. As long as they're moving things onto the platform, I'm still excited about it."
Others, such as some in the channel partner community, were concerned with a lack of a Merced commitment. John Painter, marketing manager for Computer Solutions Inc., said "the bottom line seems to be that HP will continue to support the 3000 with evolutionary enhancements, but not provide the same kind of revolutionary advancements already planned and marketed for the 9000 series. Marketing droids like myself call this 'milking the cash cow'."
Nick Demos, a longtime channel partner, said "the fact that they did not commit to implementing the full power of the new Intel/HP chip on the 3000 was not too comforting. As users grow they will need more capability and power. The 3000 must have the latest architecture implemented to achieve this."
But some channel partners have faith in HP's ability to address those needs in time. Alfredo Rego of Adager noted that the key objective isn't adoption of any particular technology.
"Given the fact that we all have unlimited desires and limited resources, HP has done a wonderful job selecting a manageable subset of desires and allocating the appropriate resources in their pursuit," Rego said. "The HP 3000 cannot be all things to all people, as some other platforms claim to be. As long as HP continues to provide, and to improve upon, the HP 3000's legendary qualities of reliability, performance and reasonable cost of ownership and operation, the worldwide HP 3000 community will continue to support it and trust it for mission-critical applications."
Waiting to see
Some customers and channel partners found HP's strategy gave them
something to
hope for and something to watch. "The thing that's missing from the plan is
something
to address new applications," said Birket Foster, chairman of the software
vendors'
special interest group. "It could have been more detailed, but it's
certainly enough to let
people know they have a solid growth path for the next five years. And some
of the
[performance issues] people are complaining about will go away because of
the higher
speed networking. Think about what will happen with Samba at 100 megabits per
second."
Rene Woc of Adager said HP's plan is a good starting point, more vision than customers are getting from vendors such as Oracle, which won't yet announce the platforms where it will support Oracle8. "What we make out of the announcements is up to us," Woc said.
Bruce Toback, a longtime developer of HP 3000 solutions and whose company OPT is a registered developer for HP peripherals, said that the evolution of HP's plans is a better match than the all-ro-nothing approach when the 3000s got a 32-bit processor.
"An evolutionary strategy, starting with a 64-bit file system, makes a lot of sense," Toback said. "We may even have the time to see it through: I'm prepared to believe that after the last couple of NT debacles, IT managers are more willing and able to stand up against the kind of "feature article in Fortune Magazine-driven" change that's plagued HP 3000 users for the last 15 years."
And one channel partner is willing to hold HP to a promise that goes far beyond its five-year plan. "The last firm promise I heard from HP was: 'If we're still using computers in the year 3000 A.D., there'll still be HP 3000s in use,' " said Wirt Atmar, president of AICS Research "Personally, I intend on holding them to that promise. I'll write back in 1003 years and let you know if it's true."