Looking Over a Surprising Year

Doug Greenup has seen the last year prove him wrong about the 3000 market’s transition movement, and he’s happy about the unexpected turn of events. Greenup founded Minisoft more than 20 years ago, a company that challenged bigger firm WRQ for emulator market share and then built up a strong following for its middleware products. Minisoft serves more than 10,000 companies with one or more of its products, according to Greenup, giving the firm a broad perspective of how the HP 3000 installed base has reacted to HP’s withdrawal from the market.

Greenup has been glad to see the customers take the prudent path of sticking to the platform while they study their options. While counseling customers to take their time, Minisoft’s staff has been at work expanding its product base during the year, releasing a new PDF forms product and adding Eloquence support to its database connectivity tools. The company has also stayed out of the migration business, despite an invitation from HP to provide such tools.

Greenup and the Minisoft team occupy a unique position in the marketplace, by our accounting — a firm with a vast 3000 installed base that sells products in other markets and is sticking with software as its core business. Minisoft supports Eloquence, but doesn’t sell it; sells cross-platform print and driver products, but doesn’t promote migration. We wanted to ask Greenup what he’s seen from Minisoft HP 3000 customers in the first year of Transition, and what path the future holds for what he calls a very smart customer base. We spoke in the week before the end of year holidays.

Minisoft is among those companies working both the migration and homesteading sides of the 3000 marketplace. What do you see among your customers regarding a readiness and willingness to migrate versus staying put and improving what they’ve got?

These customers are so smart to wait. The way they’re hesitating like they are is brilliant. HP-UX is going to be around for awhile, but now that the vendors are embracing it, I think Linux has a shot. And I didn’t think so a few months ago

I took a call a few months ago from a university that’s going to skip over HP-UX and go right to Linux. They’re in no hurry. They’re looking to migrate, and they’ve got some Linux boxes in house and are going to use our ODBC driver to link into Linux.

My belief is that the migration vendors are really overestimating their opportunity. These customers who are waiting are smart, because Linux may pay off for them. I’ve always counseled customers that there’s no hurry.

There are people doing some stuff, and people doing absolutely nothing. We see a few migrating, but they tend to be the ones going to a new vertical package. Or it’s big companies that have acquired someone, and they were required to migrate. What’s ironic is that some companies are centralizing on big iron, and some are still decentralizing. At International Paper they have no more IS department; those 40 people are gone. [Chuckles] I don’t know if that’s a migration or an absorb-tion.

The people who have home grown applications who have done a lot of tinkering with them are not moving. I think they’re still banking on OpenMPE. I think when push comes to shove they’ll eventually move. They can’t find an off the shelf package, and that’s where the migration opportunity is for an MB Foster or an MBS or a Speedware or a Lund. These guys might be an opportunity for those partners. But they’re also the customers who are the most technical and have the resource to do it themselves.

You’ve chosen to skip the business of offering migration services. What’s the thinking behind sticking with software?

Migration is a one-time thing. We talked about it because HP came to us — one of the few times they did — and said they were looking for companies that would be interested in migrating customers with tools. We didn’t think it was a long-term opportunity. You put a lot of energy into it, and then it’s done. It’s not an enduring business, no support revenue. There may be a product you develop, but could those products be sold after the fact? The answer is probably not.

Instead of doing the migration, we decided we would focus on building products that would have a life beyond this great migration. Ways to adapt our middleware so it would have some use after someone has migrated. An example of that is our ODBC and JDBC drivers can support Oracle now. We support Eloquence. So just maybe, when our customer base gets to Linux or Unix, they’ll find value in our middleware and continue to use it.

We also wanted to work on building other products we could sell on Unix, NT and AS/400s that have nothing to do with migration: document forms, archiving. We felt it was a better use of our resources, rather than having them help companies migrate, to keep them home working on products. We’re a products company, not a services company. Some of these other companies have had a lot of experience in selling bodies to do things.

We’ve made great strides, since all our products run on all these other platforms. The migration stuff ties up your best people, so they’re not building great products.

We’ve heard that Java has helped move software to other platforms easily for Minisoft. Did it do its job of getting the 3000 compliant with the rest of the computing world? How well has it worked for companies writing software as you do?

Java really came late to the 3000 platform. It’s a shame, because Java runs pretty well on the 3000. By the time Java showed up, the 3000 customer base had lost so much of its technical expertise, and it couldn’t get its arms around Java.

Java helped us tremendously. We took advantage of it. All of our products are written in Java, and we continue to build in that language.

You’re still writing software like eDirect for the HP 3000. How do you justify your outlay of expertise for a market which the vendor is leaving?

A lot of our customers ask that. It’s not very hard to do, because we develop in Java. The 3000’s Java is good enough for our purposes. Java’s promise is write once, and run everywhere, and it does. It takes about a week at most for a product we’re developing on Unix, for example, to run on the 3000. That’s not much of an investment. And the products run the same on all the platforms.

We made the investment in Java years ago. If we were in C, where we were a few years ago, it would be harder to justify the continued investment.

The other reason we continue to develop is that we have a big installed base on the 3000. We have 10,000-plus customers running one or more of our products. We’re very widely installed with our emulation package. Our ODBC driver in the last few years has been delivered to a very large number of HP 3000 shops. It makes all the sense in the world to deliver technologies like eDirect to them with the message that when the customers move, they can take it with them. You just transfer the license.

You offer products across a lot of platforms. Which alternative platforms are the favorites among the 3000 sites you’re serving?

A few months ago I would have said the 9000. When HP made its announcement there was a lot of emotion and resentment toward HP. We heard a lot of “I’ll never buy from HP again.” In the end, the majority of customers are looking at the 9000. I emphasize looking, because Linux is a sleeper. That might still mean it will run on HP hardware. For now, the safe choice and the path of least resistance is HP’s Unix. The hardware looks the same.

The next most popular platform, which has surprised some of us, is Windows. If you’ve got a gigantic database, with gigs and gigs of data, 400 users, Windows probably can’t scale. I haven’t seen companies like that adopt Windows. The customers who run low-end 3000s, we see them giving some real consideration to Windows. The smaller Ecometry shops are migrating to NT: Massage Warehouse, PetDoors, Aquascape, are all NT now. These databases aren’t that big, they have only 20 users or so. For companies like Hickory Farms, the path is the 9000 and Oracle.

Is Linux ready to replace MPE/iX this year, or is it still growing up to enterprise grade functionality?

I don’t know what the timeframe is, but there’s no question Linux will be a player. There’s a stigma still, and I just don’t see a substantial company with some IT staff, that’s not a university, has bet the farm on Linux yet. It’s still just a little bit out there. Having said that, we’re seeing a lot of Linux Web servers. We’re seeing Linux making its way into those companies for a specific need. If it proves to do well, then maybe they’ll take it the rest of the way.

But even if the technical comforts haven’t arrived yet, it sure looks like Linux is a better choice for price — at least compared to the HP 3000 cost of ownership. What have you found?

In our experience we’ve competed with WRQ on price. Our products were equivalent functionally, but there were companies that made decisions amounting to spending three or four times more for the other product. It was based on a perception that they were a more substantial vendor. Many times companies will spend more money for the security, comfort.

It’s not always price. Free comes with its own set of issues. I think you’ll find companies will pay a little more to be comfortable. Remember, people get fired for screwing up, especially in for-profit companies these days. I think they’ll be more risk-averse, and Linux has more risk. But I think its risk is becoming less.

HP’s announcement of last year about the 3000 makes Oracle more compelling for some 3000 sites, as they look at where they’re going next with their applications. How can the right database access tools make adopting Oracle easier for 3000 sites?

They can use our products in a lot of different ways: to migrate, or to do what most of them want to do, which is co-exist. Most of them don’t want to unplug the 3000, and be on a new box. In most cases they’ll have the 3000 awhile, and it will be a player along with new boxes. They’re looking for tools that will allow them to coexist.

The ODBC and JDBC drivers can go with you in a move, so there’s still value in them. We try to protect our customers’ investment, and quite honestly, our investment.

IBM says its iSeries will be a less costly alternative than Unix solutions, because customers don’t need a DBA for DB2. What’s been your experience with that kind of claim among your customers?

It may be accurate, but I don’t think the customers will listen. I don’t know what advice I can give IBM to be more effective with these customers. There’s been some talk of the iSeries among our customers, and we sell to the AS400 and iSeries market. I haven’t seen very much movement toward IBM by the 3000 community. Think about when many of them bought the 3000: they probably considered IBM and ruled them out. I think it would be extra hard for a lot of those guys to go IBM. It gets back to the change thing.

It’s not as strange for me to go to a 9000. It’s the devil you know, less change and less traumatic. If HP makes it pretty simple, they’ve got the home court advantage. IBM will get a few, but among those they’ve already gotten, IBM was already somewhere in the company.

Is Eloquence ready to take over for TurboIMAGE?

We’ve spent a lot of time with it, and it’s pretty damn impressive. If you think about it, that product has been around forever and it’s been underestimated. It works well, performance is good, and it’s easy to support. We had our ODBC driver running against it in very short order.

The problem [Marxmeier AG] will run into is the company’s size. There will be some doubt for a lot of people; Oracle is so overwhelming with its name recognition. But we’re so impressed with Eloquence that we will promote its use in our customer base wherever it makes sense. I think it’s a wonderful choice if people want the familiar feel of IMAGE and people don’t want to employ the overhead of Oracle.

What have your customers been doing in the first year of the Transition era about the HP announcement?

A lot of it has just been researching, gathering information about what’s out there. The smart ones have been investing, stocking up on 3000 hardware and components. They’re going to use this baby for another three to five years of productive use.

At first there was the denial and the shock, the anger. Then they realized it really was going to happen. I think the people who migrated in the last year were people who were going to do it anyway. I didn’t see anybody pull the plug and scream fire. I think they’re doing a lot of soul-searching, asking what applications are important to their business.

Is the economic slowdown a good thing for the 3000’s homesteading movement, since it might keep customers from jumping off to less cost-effective platforms while homesteading gathers its forces?

I think the slow economy has prevented some customers from taking action later on in 2002. Once the economy turns up, I think you’ll see some companies take action.

Anything that keeps the customer base on the box and slows down the attrition is an advantage to the homesteading group. I don’t know if the OpenMPE movement is going to succeed or not; I hope it does. I think it would be neat to have an emulator or two, something that would allow the operating system to live on.

If an emulator became available, would you have your products tested for use on it?

Of course. It would absolutely be to our advantage to have a lot of those many customers of ours stay on the box. I think the slow economy plays to that group. I hope HP will agree to do something. But if HP doesn’t, I don’t think the installed base will go away tomorrow.

How has your outlook for the 3000’s prospects changed in one year’s time from the HP announcement?

I was a lot more negative then, just after the announcement. This shows you how much I don’t know: I thought the switches were going to get flipped and people were going to be out of here.

Looking back now, that’s definitely what I thought then. But resistance to change and the slow economy have meant very few customers have moved. That’s what has startled me — how few have left. It makes perfect sense. You can’t turn on a dime. They would take the time to understand their business, what’s critical.

Do I think there will be a lot of migrations this year? I don’t think so. It’s going to be a lot slower than I thought.

How long will you be selling and supporting software for the HP 3000?

We will absolutely continue. We don’t have any fixed date. If there are customers in any significant number around in 2010, we’ll be supporting our products. We came out with a word processor in 1982 for the 3000 called MiniWord, and we still support it today. We just sold a MiniWord license today. It runs on 3000s, Windows, 9000s, even HP 1000s.

I’ve got to believe by 2010 most of the base will be somewhere else. But if I’m wrong, that’s great too. We want to continue to innovate and support and understand what’s happening out there. It’s hard, because there’s a lot going on. We’ve got a wonderful customer base that has been good to us.

I just wish MPE could be released, to offer a roadmap for the benefit of the customer base. I wish as an organization HP could be better at cutting MPE loose. It clearly is not a business HP wants to be in anymore, so just find a way to let it go.