In our interview with Allegro's Stan Sieler, we asked the veteran developer what has changed about 3000 options for the future. His answer identified a significant shift in the definition of migration. He also spoke about Allegro's own season of considering an emulator project, the tech challenges that will be outside of the system's capability, and how his practice of magic has shaped his exemplary technical career. On the occasion of his 30th year with Allegro Consultants, we spoke via iPad in November, just as the US was switching to back off Daylight Saving Time.
In the first year after HP's 3000 announcement, there were a list of options of what could happen to the community over the decade to come. Is there anything new on that list?
There are the same options but with one difference. Migration means something different now. It's not migrating your app with a 3000 lookalike shell on a Unix machine. It's migrating to Stromasys. It's a variation of 3000 Forever.
We still see people coming out of the woodwork that we've never heard of, using 918s, 928s or newer machines. They have no intention of leaving because they have no funding to leave, and now they've encountered a problem and they're reaching out to the rest of the community. We see people who tend to be on bigger machines, who are either running into limitations, or they're worried about the continued maintainability of the hardware. They are looking at high-end Stromasys solutions.
We were concerned that at the time, in addition to not yet having HP permission, that we'd face potential legal action if we did anything. We didn't want to open that door to HP. I kind of regret that now, because I would have approached an emulator a little differently than Stromasys, and I think that might have had some payoffs.
We've certainly reached out to Stromasys several times to help them with performance limitations that they're encountering with their implementation. I'm hoping that with some of the other 3000 vendors in the process, they may be able to put economic arguments in place that will help convince Stromasys to still pursue that help.
What do you think of the prospects for this emulator making a lot of difference for customers staying on the HP 3000?
I think if they can solve their high-end performance challenges, then they might be able to make some big sales to those kinds of customers. The problem: I don't know how many of those people there are.
It's true: managers are moving off the 3000, and so are moving away from IMAGE. Out of all the SQL databases you've seen, which one is the smoothest in replicating what IMAGE does for MPE apps?
Eloquence. I really like Eloquence. Michael [Marxmeier] has done amazing things with it. Tech support from him is immediate and reliable. He doesn't have problems with you publishing benchmarks. Eloquence has a lot of nice features in it. It has more features than any other SQL database — plus the IMAGE compatibility. It's a win-win situation, it seems to me.
Do you consider the 3000 has always had a tech boat anchor that made it obvious HP would leave it behind? Is it the equivalent of an unsupported system by now?
It's certainly true about CPU speed and amount of memory, stuff like that. That doesn't mean it won't run perfectly fine.
Are there a set of new tech challenges the 3000 is never going to meet, important challenges?
That would imply that this is going to be a new product you write, and nobody is ever going to write a new product for the HP 3000. If you are doing a new application, it's probably going to talk to a database. Almost anyone you hire will know how to do SQL stuff, not IMAGE stuff. It's just too far behind the times for a new application.
Of all the many projects you're worked on, which stand out at the most fun for you?
For projects, creating SPLash!. I worked with Jacques van Damme. In the very early days, Jason Goertz was helping out. But I remember sitting with Jacques in the HP Migration Center and there was a LaserJet sitting there between us. We had Post-It notes that said things like "tree building" or "generate code." Each was a name of the 20 modules that made up SPLash!. Our source code control system was that if you wanted to modify something, you took the Post-It note off the printer and put it on your terminal. It worked well because you had instant communication with the other developer.
There was that, and then our work with Alfredo Rego on repacking detail datasets. That was Steve and me working with Alfredo, and to some extent Fred White. That was something where data integrity was of absolute importance. Yet it still had a lot of opportunity for using interesting technology, doing things efficiently and fast.
How do you think practicing magic over the last 15 has had an impact on how you approach your day job?
I've always tried to think outside the box, and with magic it's easier to do. If you're developing a magic effect, you tend to look at the end result and work backwards. That the way I've done a lot of my 3000 stuff — like when I think I was the first person to propose intercepting disk IOs — I remember sitting down with Joerg Groessler and outlining how it could be done. And so basically giving him the idea for the online backup on the Classic HP 3000s. You could do it behind the operating system's back by intercepting disk IOs.
You don't start out by saying, "what can I do, and where will that lead?" You take the end result, intercepting disk IOs, and work backwards. Sometimes that's the same thing with magic. You say "I want you to be able to look at the card in your hand and see it's not the card you thought it was, but it's a different back, and a bigger card than you thought it was."
Sometimes a technique comes out for the 3000 and you think of what you can do with them. Like procedure exits came out, and you say, "What can I do with these things?"
If you could talk to the Stan of 30 years ago, what would you tell him to pay attention to?
[Laughing] Buying Apple stock. I would say pay more attention to the Internet and how to link computers together. About 20 years ago, my ophthalmologist asked me where the future of computers is going. I said the future is with computers working together. And I think that's still the answer. We're beginning to get there, but we're not there enough yet. I can't leave this iPad and walk over to my desktop, and resume this conversation yet, like nothing has changed.