How Software is Built

A blog forum to provide deep dive analysis and community conversations about software development models. For more details click here.
Tags:

Interviewers: Scott Swigart, Sean Campbell, and Richard Bowler.

Interviewee:Steve Morris

In this interview with Steve Morris the director of the Open Technology Center we asked him about:

Scott: Steve, thanks for taking the time to chat.Could just take a second and tell us little bit about yourself and the OTBC?

Steve Morris: Sure. I’ve been executive director for about nine months. OTBC has been around for a couple of years. My background is very much a technology, marketing, and management background. That started at HP, and now I’m mainly working with local high-tech companies largely in the EDA and semiconductor test areas. This includes both software and hardware.

OTBC is an incubator for start-ups. It’s funded by the City of Beaverton, related to their economic development plan, so what they’re interested in is not really pushing open source or any particular dogma;they’re interested in jobs.

The feasibility study for this place was done as a software company incubator, and it got focused early in its life on open technology.

Since then, we’ve broadened it back to being technology and computer focused more generally, because focusing on open source was just too narrow to meet our goals. The logic there is basically that if your goal is to find lots of early stage companies with a lot of scalable potential for creating jobs, then it doesn’t make sense to limit yourself to a narrow niche.

Richard: Why did it originally lean toward open source? What were the arguments?

Steve: We were very much focusing on Oregon’s cluster technologies, of which open technology is one. We’re very open source friendly, and certainly encouraging companies to use open source in ways that are appropriate to their business model. But, we’re certainly not exclusive to that, just like a VC. You won’t find VCs being too narrow, because they want to consider a reasonable breadth of good business plans for growth, and we’re much the same.

One of my beliefs in the open source discussion is that you would be hard-pressed to find very many software development projects that did not use open source in some way.

Actually, I would challenge you to point out one that doesn’t use open source in some way, because software engineers like to use open source tools. At the same time, one question I see a lot of companies not asking, that they should be asking is, "How are we using open source today?” whether it’s tools, or just incorporating code into their product. Was that a conscious decision? Did they look at the business consequences? And should indeed they be using open source in a different way, either more or less than they’re using it today?"

Scott: What do you feel are some of the reasons why open source has so much gravity? Why does it draw people to it, especially around start ups? And what do you see as being some of the reasons why very early on, when they’re wondering if they have a viable company even, they should be concerned with the how and the why of open source, and how it might affect them as their business grows?

Steve: Well, first of all let me give you a nicely documented example that’s not exactly a start up, but it’s illustrative. Ben Berry, the CIO for the Oregon Department in Transportation, decided to do an audit a year or two ago. Using some automated tools, he did a search across all of his ODOT computer systems to count every open source tool that we could find– just literally to find out what was there.

They found, if I remember right, about 3,000 or 4,000instances in the ODOT computer systems of open source tools. ODOT at that point had absolutely no policy or strategy about open source. There was no conscious management behind this.

It turned out that this was because open source development tools were free. People at ODOT could download them and use them with the zero red tape. And when you work for state government, zero red tape is a pretty attractive thing.

And so in that one particular example, it was because they were good tools, they could do the job for a software developer, and they were free. So they didn’t have to ask anybody, and they already had the appropriate rights to install the software.

That’s an interesting example of one motivation for using open source development tools, which is different from using open source code in your product. Because the tools were free, they were an easy way to get the job done, while eliminating the need to go through red tape.

Scott: On the other side, you know a software start-up company’s going to have to decide whether it’s going to go with an open source development paradigm for its software product, or whether it’s going to go for a closed source. Some choose one, some choose the other. What are the main reasons companies do choose open source as a method for developing their main product?

Steve: Of course, some people choose a mix, and that’s almost an easier question to answer, because if I’m doing a mixed paradigm, probably what I’m doing is latching on to open source stuff that exists, so I don’t have to re-invent it. For instance, if I’m doing a web application, why would I want to re-invent MySQL? On the other hand, I wouldn’t necessarily want to go off and buy an expensive commercial product either, if MySQL would work.

Richard: But in a mixed case, their main code might be closed, so they’re still going along with the old private IP route, as opposed to making a profit on a product that’s entirely open source. Building an open source component is one thing, but making the choice to develop your main project in open source is another. Do you have a sense of the business models that are behind that — the reason people choose that route?

Steve: I certainly would not claim to be an expert on that, but I think there are a couple of different reasons. One is that there truly are people who believe that software ought to be free, or close to free,as in beer, and it really is a philosophical thing.

There are others who are going to use an open source paradigm, who really are thinking more pragmatically. They see something that already does a lot of what they need to do, and it’s open source, so they can use it, and it gets them where they need to go with less effort. People with that more pragmatic attitude are obviously looking at a monetization strategy that’s different then selling source code.

For example there are lots of ’software as a service’ companies, which really is essentially a web based service. If you can put together most of your software by leveraging something that’s open source, that can get you to your software as a service play a lot faster.

Scott: One thing you said was that some companies start grabbing different open source components and tools, but they don’t necessarily put a lot of thought into it. And you alluded to the fact that it might have business implications down the road. Can you expand on that a little bit?

Steve: Take the ODOT example, where they had software engineers simply downloading it, and using tools with no license review,no policy, and no processes. One interesting question is whether they are going to download something and use it in a way that might violate the license.

My guess is, when you’re talking development tools, that’s probably a pretty rare problem. It’s a lot different, obviously, than downloading some code and putting it in your product, in which case you absolutely can have some license implications.

Scott: In the companies that you deal with, have you seen it happen that people find something that does what they need, and they bundle it with their product and then realize that the licensing terms are going to have some ramifications that they don’t necessarily want? I know that’s a concern, but I’m curious how often it really gets that far.

Steve: It gets that far a lot if you have a mixed model. If part of your product is open source, and part is proprietary, you have to be particularly careful from a licensing standpoint to build the right wall between the license for the proprietary stuff, and the license for the open source stuff. If you’re not careful, your open source license can infect your proprietary software.

Scott: I know that Microsoft cannot accept any community written code for anything that goes into any of their products,because they’re terrified that somebody will do a Google search, find a section of code and say, "these 10 lines of code were originally under the GPL license. Therefore, Windows has to be open source now." Maybe I’m exaggerating a bit, but it’s something like that.

Do you encounter companies that go the other direction? Do you encounter companies that just say, "We’re just using 100% the Microsoft stack: Windows, .NET, Microsoft SQL Server. Or, from your experience, is that pretty uncommon as well?

Steve: I certainly know a few companies that wouldfit that description, although I don’t think we have any that quite fit it at OTBC. My impression is that Microsoft represents such a huge market that you get a lot of companies out there that focus pretty exclusively on a Microsoft environment.

Scott: Do you find that companies put a lot of forethought into it? Is it because the company founder has a certain preference because of their experience, or that they pull in certain developers who lobby really hard for building on top of Linux, or building on top of Windows? Do you see it more often being a really conscious, up-front decision, or just something that kind of happens?

Steve: I see it being a pretty conscious decision most of the time, because it typically comes down to what your target market is, and what platforms you have to support. If your target market is largely Microsoft-driven, that guides you in that direction.

On the other hand, if your target market includes a fair amount of Linux, and Mac, or whatever else, then clearly you’re going to think about it differently. Most software companies are pretty pragmatic about who their target is. Obviously, their choice of platform and development approach can eliminate potential customers, and so they need to think about that fairly carefully.

Richard: You said that over at the OTBC, it’s not specifically or explicitly all open source anymore. What percentage of the companies that you have in residence there are working on purely open source products?

Steve: Well, including open source products with a software as a service business model, I had at in at least three or four,probably more.

Richard: Can you expand a little bit on what that model looks like?

Steve: One of the companies is called"GoSeeTell." They’re building what is essentially a social networking play having to do with travel. They’re taking an interesting approach where they target a particular sponsor. So, for example, they’re working with Travel Oregon, putting together a web site for Travel Oregon that leverages comments and suggestions from people who have actually been to the places.

But, at the same time, they’re doing one for a Fortune 500company, which is targeted for that company’s employees. It leverages the same database of user feedback as the other portal does, but it’s a portal that’s customized to some specifics that are of particular value to the company’s employees.

It’s very much in a social networking play in that it lets you look at suggestions from other people that have been to a place. And you go beyond that to entering your profile and having the system start to make recommendations: "People like you really enjoyed this place," and then you can hear some of their comments.

The infrastructure is all built on open source tools — things like databases, content management systems, and so forth.

Richard: So, they’re using open source behind the scenes to do their development, but their actual product is really a service offering through the Internet?

Steve: Right. The product itself is a service, and but some of the underlying software is open source.

Scott: In the companies you work with, do you see that as being a primary model of open source, where they’re building their own intellectual property and their own proprietary things, but they’re building it on top of open source because of the advantages that they see open source having?

Steve: It’s mainly what we’re seeing here now, at OTBC at least, in that the majority of the companies we have here are software as service pledged, and to one extent or another, most or all of them are leveraging open source, some pretty aggressively. I don’t know how representative our little sample is, but it certainly is very much the model that we’ve got going on.

Scott: In the Portland metro area, open source is pretty strong, but I would guess that it’s not unique to this geography. The reasons it makes sense here are geographically independent.

Steve: And when you’re building software as as service, again, that’s one place where it makes sense. If you can get pieces that are already built and that you don’t have to pay a royalty to use, then that’s great. And when what you’re offering is a service, you’re pretty safe,in terms of not violating anybody’s license.

Scott: In cases where they’re using a hybrid approach, where they’re not just going 100 percent open source for their infrastructure, what are some of the examples of where they’re choosing a closed source, proprietary component to be a piece of it, and what are some of the reasons you see behind that?

Steve: One particular company that used to be here definitely had a mixed strategy. They had some basic tools that were open source, and then they had some more advanced tools that were proprietary.

The business model was definitely based on proprietary software, and they’re using open source items as marketing hooks, getting people aware of the company and the capabilities. That’s certainly one model you see in the mixed mode — providing some value through an open source solution that will get your name out there, get users involved, and give them a reason to find out more about your other stuff, in order to up-sell them on your proprietary line.

Scott: So in that example, they offered an open source version of their product, but there were certain up-sell features that were for-pay that weren’t open source?

Steve: Exactly. In their case, they were offering a couple of fairly fully functioning modules which were very nicely expanded with capabilities in their proprietary software modules.

Scott: I’ve seen that model quite a bit, too. Some of the things we’ve run into, like SugarCRM, certainly have a model like that,where they give a certain amount for free, but then other features — fairly essential things — are for-pay. Other times the whole product is open source,but what you pay for is support. If somebody goes with a completely open source product, it seems like support, a lot of times, is what they end up charging for.

Steve: I think those are the two basic models.You’ve got to monetize it somehow. You know the Red Hat approach, right? You’re primarily paying for support and installation tools, which is not too different from the traditional model of having a basic product that’s really cheap. In this case, ‘really cheap’ is free, but that’s obviously a model that’s been out there for forever.

Scott: Of course, it’s not unique to open source.Take a look at Adobe. Everybody can download and run the Adobe Reader which work on every important platform. But for a long time, the tools to produce PDF files were something you had to pay for. Now PDF has become so ubiquitous that the format itself is pretty much open source. You see that a lot, where companies will give away a tool that lets you consume something, but the tool to actually produce it is something that you have to pay for.

Interesting to see how that’s evolved. Even companies that are building on top of Windows seem to be doing the same kind of things. I wonder if you’re seeing any of that, where companies are choosing a proprietary stack, but they’re still going with a product that’s modular and they can build a community around, and some of the modules or add-ons or things like that are community-built.

Steve: I haven’t seen a lot of that here or around here. I see what you’re saying, and I think there are certainly things like that out there. I just can’t think of too many examples off the top of my head of them.

Scott: For the companies that go with the route that you talked about, where it’s, "Give away one version, up-sell to another version," it seems that building a community around the free version is pretty critical. Do you see secrets to success in doing that, in getting your free version to achieve critical mass?

Steve: Getting the free version to get critical mass, getting some initial users, can certainly be an effective marketing strategy, whether or not you build a community.

To me, I guess there are two different things that you just talked about. One is having something out there that’s free. As you pointed out, Adobe Reader was free, and it wasn’t open source. It was very definitely a market development strategy, much like some open source plays are, which is different from putting it out there as open source specifically to try to develop a community that’s going to actually help develop and extend the product.

To me, those are two different goals. Maybe I just want to put out something that’s free, and maybe I want to call it open source so that I’m open source-friendly to appeal to the community. But I might not really be looking for a developer community. I might put the source out there, but I might not really care. I might mainly care about people using it.

Richard: Are any of your resident companies leveraging that sort of community? Are they getting development help outside,from a community that they’ve put together through open sourcing?

Steve: Right now, I think the answer to that is no.And again, it’s primarily because the software companies here are largely focused on the software as a service model. I think you’ll see that happening in the software as a service model, although I haven’t seen it here a lot yet.

You’ve got the same rationale, even if your software is available as a service. There’s potentially some leverage if you can open source it and get some additional functionality. It’s a little different, obviously,with software as a service, because this stuff is running on your servers, as opposed to the customers’. And so you care mainly about making sure that it’s going to run well.

I have seen some companies playing with models having software as a hosted service, where the software is open source, and as another option you can run it on your machine. That’s getting closer to a traditional open source model. You can go through the cost of hosting it yourself, or you can buy it as a service on the company’s server.

Scott: If I’m a company looking at releasing as software product, if I release a portion of it as open source, I will get a broader customer base potentially because there are certain people out there that really only want to run stuff that’s open source. Even though they may never look at the source code or do anything with it. Is that a fair assessment?

Steve: I think there’s some truth to that, although I think you can add that you can also position it as an insurance policy.Companies have various types of insurance that they hope they’ll never use.Having access to the source feels like a kind of insurance.

Scott: You know, one of the things that we’re looking for are some real world examples of companies that have been “saved by the source”. Our search hasn’t been extensive, but we haven’t really found any examples of "If I hadn’t had the source, I would have been totally hosed."

Steve: I’m not surprised to hear that because it’s a value proposition that you hope you never have to use, and that’s probably how most people think about it, although I’m sure there are some people that like to buy open source because they just can’t wait to go in and tweak it.

But I’m not sure that I can point you to a specific example of somebody who just really had to tweak something for their purposes.

Richard: So far, we have really been focused on open source consumers and mixed models in terms of using open source tools to put together interesting companies with different sorts of business models.Traditionally there’s been a certain level of antagonism between the open source and closed source community. And some of the things you’ve said imply that, on the open source side at least, that antagonism is changing into something else, some level of collaboration. Would you say that’s a fair assessment?

Steve: I’m not sure. It’s certainly a continuum;there are certainly people out there in the open source world that I would callfairly "purist" who truly believe that software should be free.

On the other hand, there are some more marketing-oriented pragmatists who look at open source as merely being one of the tools in the tool belt that you use either for a marketing purpose, or for a more efficient product-development purpose, or for an additional set of promises you can make to customers.

I think there are a lot more shades of grey out there. I suppose it is changing, but changing by virtue of more folks looking at it with a fresh view and being maybe a little more pragmatic as opposed to philosophical.

Richard: Well, if you think about a bell curve and you think of people who wouldn’t touch open source on one side and people who wouldn’t touch anything but open source on the other side, it seems to me that the center is broadening. That more and more of the majority of people area gnostic about open source versus closed source. Rather they’re looking for what works best for the business model or the support that they want to put together.

Steve: Yes, I absolutely agree. And I guess in my mind the thing that a lot of people gloss over when they talk about open source is that there are these two fundamentally different ways of using it. On the one hand, using tools that help you do your software development, like tools for tracking and reporting bugs, that kind of stuff–stuff that doesn’t go into your product, versus the stuff that does go into your product.

Software engineering tools are where you’d be hard pressed to find very many companies that don’t use open source to a certain extent,because there are just so many good tools out there for software developers.

Scott: I think big companies — the Suns, and the IBMs, and the Microsofts — depend on that. They depend on an ecosystem that’s going to fill in the gaps, and sometimes that ecosystem is a third-party vendor, and sometimes that ecosystem is free and open source software.

Steve: Exactly. Some of it’s third party, and some of it’s not. Some things develop simply because everybody needs them, there’s no good proprietary solution available, and there’s no particular competitive advantage would come out of building it. That’s where you get stuff like Apache. A bunch of people needed it, so instead of everyone inventing their own wheel they all decided to just cooperate.

Scott: Steve, thanks for taking the time to chat.

Posted by campsean on Monday, December 3rd, 2007


You can follow any responses to this entry through the magic of "RSS 2.0" and leave a trackback from your own site.

3 Responses to “Interview with Steve Morris - Director - Open Technology Center”

Post A Comment