Teaching Lessons with a Long Lifespan

Frank Smith still has plenty to teach that 3000 customers can use. The founder of Alden Research, Smith started in the 3000 marketplace as a consultant in 1976, eventually developing an application to manage fleets of vehicles for utilities. But after two decades of business supporting and enhancing AR Fleet 3000, he got a calling that changed his life with the HP 3000. HP asked him to teach its customers fundamental and advanced MPE/iX skills.

Smith had found a new career in the 3000 market, one that ultimately led him to write the training courses still in use on the Web and at HP’s Education Centers. Now that the vendor has announced its departure from the 3000 market, Smith is carrying on with training, both on his own and for HP when it can assemble a class successfully. He’s also branched out to teach skills for a new HP application under Unix, Linux and Windows, finding a way to expand his experience while retaining the value of his MPE roots.

Smith is also starting a new service to take customers’ VPlus application interfaces and move them to a new platform-neutral interface from AICS Research called QCForms. (The technology, which Smith will use to expand 3000 app functionality while making programs less platform-specific, will be covered in a NewsWire feature next month.)

Maintaining MPE skills, and transferring the market’s knowledge to new staff, is becoming a critical issue in a homesteader’s drive to remain on the platform. We wanted to talk to a teacher of the HP 3000 about the available resource and the need to study a platform whose history is longer than the age of most college students. Even if the 3000 isn’t changing much, Smith still can profess a lot of value in the platform, value that must be protected through advocacy and training. We spoke with him via phone at his Maryland-based headquarters in January, at the beginning of the first year HP did not sell a 3000 — but a year when the majority of its MPE customers still must manage their need for skills on the platform.

How did you get started training in MPE?

It was almost a miracle. I had been doing development and support of AR Fleet for many years when I received an e-mail from HP asking if I’d be interested in teaching MPE. I thought, ‘This is really not me. I’ve got this prosperous business going, and I was doing support for my customers. But I decided to answer the e-mail in an offhanded manner with one paragraph of my background. I was surprised and shocked that they asked me to do a phone interview, and a couple of days later I interviewed at HP’s Rockville, Md. education center. A week later I audited an HP class, and began doing something that I discovered I loved immensely — teaching.

I severed ties with the remaining accounts I had — they were winding down their use of AR Fleet anyway — and went into full-time teaching in 1995. I’m still teaching to this day, though not as much as I’d like. I’ve developed course material, and am teaching MPE and a Predictive Support replacement that runs on everything but the HP 3000 called iSee.

What are you teaching in MPE these days?

I’m doing onsite classes for HP. The last one was for the California Department of Corrections, kind of a custom class based on the hands-on MPE/iX 7.0 workshop I wrote and delivered for HP. That was the last one I wrote for HP. I also completely modernized the Performance Tuning class.

At its peak, how much training were you doing in HP 3000 skills?

I taught at every HP Education Center, and I was busy 75 percent of the year. I enjoy helping people, and I really enjoy the instant in time where the light goes on at midweek, and they start speaking MPE and doing MPE.

Have you seen HP winding down its 3000 training engagements with independent teachers like yourself?

No question about it. The amount of work I’ve done is down to a tenth of what I used to do. I’ve developed my own HP 3000 course material and delivered it onsite to a couple of customers.

People who own 3000s worry about the skills disappearing from the market. Should they?

In the short term, definitely not, because it’s a buyer’s market out there now. There are quite a few good people out there who can’t find work. In the long term, [HP 3000 sites] have to support the community that will support them. I think they will rise to the challenge.

There are people who think that HP is out of the training business, and essentially they are. But I am still doing onsite training in the 3000 for HP. They offer quite a few classes, but cancel just as many. The delivery rate is far lower than it’s ever been.

There are some customers that are looking for training. Most recently I had one approach me. When I asked them how they found me, they said “Adager’s Web site, and since HP isn’t training in the 3000 anymore, how about coming down and teaching us?”

People must pursue training. They are losing some skills, but they are still going to be in business. So they need to do the training of folks who can keep them in business.

By current estimates more than half of the 3000 community will still be using the systems through 2006. Where do you believe the customers of 2006 will get training for their HP 3000 skills?

I, and other instructors like Paul Edwards, will still be around. I think they should look to proven experience. I really intend to continue with the HP 3000 so long as the market is there.

What do you think of the prospects for a viable 3000 community beyond 2006?

We’re a remarkably loyal group of people. Sometime we stick with something longer than we should. I think we stuck with HP’s promises for a long time and demonstrated our loyalty. I think the community will support the people who support it. These people are trusted, like Alfredo [Rego of Adager] and Wirt [Atmar of AICS Research]. These are the people they can trust; they just have to make a living for them. It’s not really difficult. If they intend to stay on the 3000, they will.

I believe there will be people on the HP 3000 well after the 2006 deadline, after HP departs the scene. I have three Classic HP 3000s in my office. Haven’t fired them up recently, but they’re all working. The hardware is all there, and applications built on those machines now run on the PA-RISC generation of HP 3000s. I sincerely believe there’s still value in that platform, even if it goes no further than it already has. Moving off the 3000 will be an economic hardship for so many people that I don’t believe they’re going to be moving.

The folks that are going to move are those who are being moved off by their application vendors. The rest of the folks — people who have developed proprietary code — these people are staying. I just did an onsite training class at the Department of Corrections in California, and those guys are developing. The have a humongous investment in 3000 technology. They’re happy with it. They’re still working with new applications with the HP 3000, despite HP’s announcements that they’re going away. They’re about to upgrade to all A-Class and N-Class machines. They’re girding for the long haul. There’s no reason to stampede off the platform, unless it can’t do something they can think of.

But don’t you agree that the homesteader’s budgets have been tight for a long while?

We’ve had three years of an economic downturn, and these customers have been extremely conservative. Prudently, I think. As far as people being stampeded off the 3000, I think all the surveys confirm that’s not the case, they’re not going to be spending money immediately. Again, I think prudently.

OpenMPE has represented some hope for these homesteaders. What do you believe the organization has to get accomplished?

I think they need to nail down HP and get hard commitments. I’m not so sure about the emulator for the future. We do have issues outstanding with the diagnostics, which will be absolutely essential for both the self-maintainers or the third-party maintainers. There is the issue with documentation, and finally the issue of SS-CONFIG, the ability to change 9000 hardware into 3000 hardware. Those issues have not been addressed in any fashion that would make me feel warm and fuzzy about depending on HP’s promises.

I’d like to see HP say, “This is what we’re going to do, and how we’re going to do it.” I’m sure it’s a question of the bureaucracy within HP. I’m sure they’re good folks. But we need commitment on those issues. Given the availability of hardware to be converted, with licenses and all the rest, that would be a revenue stream for HP. I don’t know that an emulator would be cost-effective.

You work and train in the Unix environment. Is there a non-3000 environment that is most complementary to MPE today? One that a company would be better served to cross-train its 3000 staff in?

I can’t think of one. I won’t pretend to be an expert in proprietary platforms; I haven’t dealt with the AS400 and iSeries. I would have to say, again, the HP 3000 performs magnificently.

Now as a developer and programmer, I like Unix. It’s fun. You tinker a lot, but that isn’t businesslike. Linux is the same way. Fine for people like me who like to tinker. But not a business environment.

If you want a stable, productive environment, that is a low-cost one, the 3000 is your best choice. One of the secrets that the 3000 folks don’t want to reveal is that to become a competent 3000 developer or administrator, you don’t need the breadth of knowledge that an HP-UX person has to have. It costs an IT outfit a lot more to field Unix or Linux boxes than HP 3000s. That’s one of the virtues of the HP 3000. You don’t need an army of people to maintain a single host. One person is able to maintain many HP 3000 hosts. The ratio just doesn’t work that way with Unix.

But we might be preaching to the choir here. The value proposition is that it works hard, you don’t have to spend a lot to maintain it, and we don’t have the training costs and high priced administrators necessary. Because the 3000 environment has been written rationally, for business people.

Migration advocates can rationalize that moving away from the 3000 can be a good thing, if the new platform does things that MPE cannot. Is there a serious list of business computing tasks that MPE is ill-suited for?

For someone who is as die-hard an HP 3000 person as I am, I find it hard to conceive of something the 3000 is ill-suited for. Something came to mind, though, after a few conversations and while I was rewriting class material that touted the client-server model. We concluded that client-server is a very fragile business model. I remember that someone wrote, “With client-server we build a Web of increasing fragility.”

About the only thing the HP 3000 doesn’t do well is client-server, and that’s cause for celebration. That is not high-speed, online transaction processing. That’s what the 3000 is built for. It’s rock solid. And did I mention IMAGE? With the combination of a machine that’s been up and running for so long doing high-speed processing, and IMAGE, you have a machine that’s eminently capable of most every business solution you can entertain.

MPE struggles to be a Web server, primarily because of its bandwidth. That’s not the fault of MPE, that’s the fault of crippling the systems. I’ve run Samba and Apache on MPE boxes for years, and I’m quite impressed. Certainly, there may be environments where the number of hits on the site are just too much for the 3000 hardware. But that’s primarily because we don’t get a full-speed A-Class or N-Class machine out of HP for the 3000.

Those machines are the reason I thought HP was going to stick with the 3000 forever — beautiful machines, and the effort HP spent fielding those machines.

So these servers prove that even good engineering can precede a move away from a product. Can a company gain security for 3000 homesteading by moving its application interfaces to a platform-neutral environment — even if they don’t have any plans to migrate?

Absolutely. I couldn’t agree more. There is a history, and not just with HP, that vendors move away from products. That product could be middleware, and I guess I would place VPlus in an antique form of middleware, or it could be an operating system. As we make our applications independent of those business decisions we have no control over, we do have security. We should minimize our use of intrinsics and proprietary features of languages, so we can feel reasonably insulated from business climate changes from our vendor.