X
Tech

Day 4: Borland's Ted Nielson says: "DevCo" will continue JBuilder tradition on Eclipse framework

Ted Nielson, CEO of Borland announced this morning that "DevCo" (the code name for their IDE spin off company) will work closely with Borland and be a very viable effort; they just need to find the right partner.
Written by Ed Burnette, Contributor

(Ted Nielson, CEO of Borland gave the Thursday morning keynote at EclipseCon 2006. The subject was "Eclipse Lessons for Enterprise Development".)

Burnette_DSCN0724-2.JPG


Ted: First let me get this IDE spin off question out of the way and explain what's going on. At Borland, we discovered we were doing a lot of things but nothing really well. We were investing here and there, some lifecycle stuff, some IDE stuff, etc.. So we decided it made sense to separate the businesses: let the IDE business focus on their customers and let the ALM business to focus on theirs. Borland is looking for a buyer for the IDE business. I fully expect the IDE group (known as "DevCo" internally) to work closely with Borland on an ongoing basis.

(David I., core team at "DevCo" came on the stage:) We're committed to the developers. All our customers that use JBuilder and love it should know we can spend more time on it by building on top of Eclipse.

Burnette_DSCN0742.JPG
 

We're going to be partners with Borland, and we have lots of customers and partners in common with them. We'll collaborate, but our laser-sharp focus will be on developers. We're not going to take the developers out of development. We're going to be here for a long time, not subsumed into some huge company. Stay tuned; it'll take a few months.

(Ted:) I'm not looking for a 'Lex Luthor' that will kill everything and ruin DevCo's business. It's in everybody's interest to drive the IDE properly going forward.

Open source is going to have a long lasting impact on the way development is done. When you look at the world today, we like to put things in very clean buckets, but that's inappropriate. I've worked on both sides of the fence at Microsoft and at the Java world.

If you look at the "Closed" development (software developed in private) they try to involve the community but ultimately the buck stops at Microsoft. If you then look at the "Open" approach (software developed in public), you can see it has the benefit of influencing where things are going, but it's murky who exactly are the decision makers.

Next if you look at the "Proprietary" world (code owned by a commercial company), it's as open or portable as they want to make it. With the "Open Source" approach, code is owned by the community (publicly available).

People lump Closed and Proprietary together, and they lump Open and Open Source together, but that's not always the way it works. We should look at what's the best way to develop and deliver product. Today we're seeing a blurring of markets and models. Despite what the CEO's might say, open source is infiltrating the enterprise because it lets developers get their job done better.

The IDE business is a good example. We did an inventory of how many open source projects were used in our IDEs, and 45+ different OSS products were distributed with it. The whole idea of leveraging the value of open source and blurring those lines is what we're saying should be done. The model going forward will be a symbiotic relationship.

When I was at MS in 1999, were were about to ship Windows 2000. We put a server up on the web, and challenged people to bring it down. They brought it down, the bug was fixed, it was put back up, back and forth. Then somebody said "Wait a minute, we're helping them!". For three days, there were no hits on the server. Then Brian Valentine said "The entire web couldn't bring it down for three days". Thirty minutes later, Boom!

Going forward: how do we blend the best attributes of these models?

"As open source begins to look more corporate, corporations themselves are looking to adopt and adapt more open source practices" - The Economist


When you read the press the emphasis is on "it's free!". But that misses the whole point. Side story about Richard Stallman: I was the guy at MS stuck doing the weird things. At a forum at the University of Colorado, there was a debate, "Bill Gates: Hero or Satan"? It was me, Richard, and Roger Ebert (thumbs up!). I learned a couple things. First, being an unknown guy at MS I wasn't taken seriously, and second, when Roger comes out it's a love fest. When Richard came out it all all over.

If you look at the business models, open source has become trendy. Saying you're 'open source' you can double your valuation. The business model is evolving and we haven't quite found the sweet spot. But we can apply many open source lessons to businesses now.

It's not about money, it's about a better process. OSS is a software development process that values facilitates and enforces the lessons of transparency, accountability, and collaboration. What does that mean?

Transparency covers things like making bug lists public, and project health dashboards. Most commercial CEOs would shudder at the thought of showing their bugs to the public because it would acknowledge they have bugs! The OSS community has done a fantastic job at this.

The second lesson is accountability. Software developers care more about respect, ownership, and reputation than money. Recognition is valued highly. The OSS community do a great job of respecting people that do great work (and castigating people that make mistakes). Within corporations you don't see that as often.

Of course some times this can get carried away. Like "Do not eat" labels on drums of chemical goo.

The last lesson is collaboration. The OSS community facilitates collaboration between individuals, vendors, customers, and academia. The world was saying that there was no way this was going to work but it did, and has changed the way software is developed.

What impact has OSS had on Borland? We're evolving our own internal development processes. We're developing products and processes that use OSS, and leverage the technologies. Go to the Borland developer network, and you can see many examples of this at work, like the open bug database (though we don't expose the developer names).

We're working to enable more transparency between development and management, and on improving collaboration. We're trying to get rid of the management hierarchy, do collective code ownership and integrative peer reviews, and expand centers of excellence so customers can come in and sit next to the engineers.

On the product side, we're part of the GMF project and co-leader of the proposed top-level modeling project. We're taking our own products and applying these lessons ourselves. For example, project Peloton takes the best of JBuilder such as 2-way editing, peer-to-peer, web services enablement, etc. and builds that on top of Eclipse. This lets us focus on areas where we can innovate.

Another example is Project Gauntlet. Gauntlet is a small company (4 people) that we acquired this week. It's also a product on top of subversion, with the goal of making software development easier. It has a focus on accountability, monitoring, quality, progress management, etc. This will be shipping in late 2006 or early 2007. (That sounds a lot like IBM's Jazz product. --Ed).

Finally, going forward, Borland will be using the open source community as the system software provider of choice. We're going to build on top of the core OSS systems. Today we're using Eclipse for Together, CaliberRM, StarTeam, and others. OSS will be the default foundation for our ALM products in the future.

Since I'm going to be a customer of OSS, here's what I need from you, the community. First, continued leadership. I'd like to see a longer term roadmap of what things are going to look like. Help us by providing greater insight on what the next 24 months is going to look like. Further, I don't need the "Home Depot" of software, I need to know how things are going to integrate, how projects are going to work together at top levels, and so forth. This is necessary in order for a vendor like myself to build on.

Second, there's a mandate of diversity. The best software is built when it's refined by fire. Having just one vendor focus on something doesn't assure long term success. (This multi-vendor requirement was mentioned again in Mike's Callisto love-fest. --Ed)

Finally, it's all about quality, quality, quality. Continuing a focus and emphasis on quality is very important. Please let this drive your processes going forward.

Take-away quote: "You've done a fantastic job. You're going to leave the software community and the world in a better place than where you found it."

Editorial standards