In November Borland Software Corporation announced the formation of CodeGear, a subsidiary company set up to sell its developer tools. The move came months after Borland announced it would sell the division to focus exclusively on the application lifecycle management(ALM) market.
I caught up with Asia Pacific Director of Technologies and Evangelism, Malcolm Groves, who is at the helm of the new company in the region to talk about the new changes and direction of CodeGear.
First up, why has Borland made a separate company for it's IDE business?
Borland has grown into a company with over 14 different product families, covering every phase of the development lifecycle. However, many of these products vary greatly in how they are sold (eg. Direct, partners, ecommerce, etc), who we sell them to (from students, independent consultants, small business all the way through to corporations), how we support them, how long the sales cycle is, etc. It is quite complex to balance all of these differences efficiently.
Further, the two broad product categories, Application Lifecycle Management (ALM) solutions and Developer Tools, really should not be tied together. By this I mean that the ALM solution really should be IDE-agnostic, eg someone who uses Visual Studio, say, should be able to use StarTeam or Caliber. Equally, someone using Jbuilder or Delphi should really be able to use it with a non-Borland ALM solution. While possible as a single company, it is much easier to pursue these
opportunities as separate companies.
The solution of separating out the two businesses makes a lot of sense to me. Borland can now be fully focussed on the things their ALM customers are telling them are important, and equally CodeGear can be 100% focussed on the needs of Developers. Customers of either product set deserve no less than the complete attention of their vendor.
What products are now under the CodeGear banner?
The current product line up for CodeGear is:
* JBuilder, including the just released Jbuilder 2007
* The Developer Studio family of products which includes:
- The Turbo versions of all of the above
* Interbase - our high-performance, low-admin SQL database
So how will things be broken up locally? How many staff will CodeGear have in Australia and where will you be located?
At the moment CodeGear has 7 staff in Australia and New Zealand, located in Sydney, Melbourne and Auckland. Our business is extremely partner focussed, and with our distributor Softgen, we're taking steps to increase this partner focus further. This means the local staff mostly deal with marketing, sales management and technical evangelism, while we
rely on our family of partners as our extended sales team.
How do existing customers contact CodeGear for support?
The only difference in our support that our existing customers should see is an improvement. The existing tech support agreements still stand, the phone numbers have not changed, etc, so on that front there should be no disruption. The plus side is that the number of technical people in the region focussing on these products has increased quite dramatically, including people like Tim Jarvis and Richard Vowles here in ANZ who are part of our Technical Evangelism team, and Chee Wee Chua, part of our Tech Support team in Singapore, who is very active with our user groups here in Australia.
What will happen to Borland's technical resources for their IDE business on the Borland Developer Network?
The BDNTV vidcasts and BDNRadio podcasts, the CodeCentral online sourcecode repository, QualityCentral for bug reports and feature requests and all the other applications and resources that make up Borland Developer Network are now all part of CodeGear. David I, John Kaster, Anders Ohlsson and the rest of the team who manage BDN are all part of CodeGear and already at work on new projects to provide better access and information to our community.
Last year Borland's chief process officer, Bill Curtis told Builder AU "Everyone thinks of Borland as a provider of tools for individual developers, and they are really good at that, but that work is going away because of open source, Eclipse and other things." If Borland don't see much commercial gain in making developer IDEs then why start CodeGear?
This may be Bill's personal view, I'm not sure, but it's certainly not the view of CodeGear. I don't buy the argument that Open Source has taken away the need for companies like CodeGear. CodeGear's job is to provide developers significant value over free and Open Source tools. Any vendor who fails to do that will not compete, but we've been successful for 23 years in this space, and the whole time we have competed with free alternatives. Jbuilder is the right strategy, I believe. We should embrace Open Source, then spend our resources building extra capabilities on top of the Open Source base, rather than waste effort duplicating what Open Source provides. Open Source is an opportunity, not a challenge.
Besides which, do you really think if you asked developers they'd tell you that the tools they use are perfect, no need to add anything else, thank you very much? I'm sure you've heard the quote that is supposed to have come from the head of the US Patent Office in the 19th century. He's alleged to have said "Everything that can be invented has been invented". To me, this sounds the same. The thought that we've solved all the tough development problems is ridiculous. We at CodeGear have got enough ideas about how we can make life more productive and fun for developers to keep us going for a long time. Anyone who thinks that it's game over has simply run out of ideas.
This week CodeGear announced it's first official new product — JBuilder 2007. What's in the new version you think will excite developers?
This is the most feature rich version of Jbuilder for years. This new version is based on Eclipse, and we've done a lot of work to make sure previous Jbuilder users can automatically migrate their projects across. Of course, being based on Eclipse also means that they can now access the massive collection of third-party Eclipse plugins.
However, I think the Eclipse base, while important, is actually the least interesting change. We've added support for EJB 3.0, including the ability to automatically convert descriptor-based beans across to either annotation-based or x-doclet based beans. We're also supporting a whole range of EJB 3.0 servers, such as Glassfish, JBoss, and others. We've
added visual EJB and Web Services designers which leverage the LiveSource technology inside Jbuilder. This means you can change the model of your EJB or Web Services and the source will be updated automatically, and vice-versa. There's also an all new version of our profiling tools, covering memory, CPU and JEE profiling, code coverage, thread debugging and deadlock detection. We've added full UML modelling capabilities, including the ability to automatically generate Sequence diagrams from your sourcecode, sourcecode audits and metrics, and on and on.
I also haven't mentioned ProjectAssist and TeamInsight. Most development environments, if they integrate with version control tools, bug tracking tools, project management tools, etc, typically have vertical integrations. By this I mean that while there might be a plugin for your version control tool, it usually doesn't communicate with the plugin for your project management tool, or your bug tracking tool. While useful, this type of vertical approach to integration is missing some
opportunities. TeamInsight meanwhile, takes a horizontal approach to integrating these types of tools. So for example, you can have a single view that shows you all the tasks you are working on, wherever they come from (ie. bugs in bug tracking, features in requirements or user stories, etc). That's what I mean by horizontal integration, and that's an example of the kind of capabilities TeamInsight provides. At the moment TeamInsight supports an open source back end, specifically
Bugzilla, Subversion, Continuum and Xplanner, however we'll be adding TeamInsight support for Borland, Rational and Mercury ALM stacks, as well as more open source, in the coming year.
The new JBuilder is built upon the open source Eclipse platform. Do you think CodeGear will follow this trend in the industry and build new tools on this platform?
We've been incorporating Open Source technologies into our tools for many years, from unit testing frameworks, web and application servers, open source ALM tools, and component frameworks, just to mention a few, and I'm sure this will continue. While I'm not convinced it makes sense that all our future products be built on Eclipse, I'm sure JBuilder won't be the only place we leverage it.
As CodeGear moves on with it's product roadmaps what are some of the innovations you think the developer community will be excited about with upcoming releases?
The first half of 2007 will see a new release of Delphi, which while adding support for .NET 2.0 and 3.0, continues our commitment to native Windows development. Unlike some of our competitors, we recognise the need to continue supporting Win32 developers, and in fact are continuing to extend both the language (eg. Generics for Win32) and the IDE for this platform. Further out, we're adding 64-bit native support and as always, a very easy migration path for people to move their 32-bit code forward. On the .NET side, ECO, our model driven framework, continues to be extended, with support for VCL.NET being added, while we're updating the Delphi for .NET language to support the new goodies in .NET such as generics, nullable types, etc.
Interbase has just had a new release that adds a bunch of durability features, such as Journaling and Point-In-Time recovery, and performance features such as batch updates, new query and OLTP optimisations. However, next year it will be getting some cool new features to make it much easier to embed in the applications our customers build, including
DLL deployment and mobile deployment.
There's more to come too, as CodeGear wants to help all developers, not just Delphi, C++, C# and Java developers. I think I've mentioned to you before our interest in things like PHP, Python and Ruby. I've been spending quite a bit of time lately exploring Ruby, and there's a lot to like there (although I confess I was a bit of a Smalltalk fan at university, so maybe I'm biased). No announcements yet, but if you were to eavesdrop on the conversations going on within CodeGear, you'd get a
lot of hits for these names.
Anything else you'd like to add?
I've said this to him before, but I want to publicly thank Tod Neilson for making the decision to let CodeGear run it's own race. It's the right thing for Borland, the right thing for CodeGear and the right thing for all of our customers. I've never seen a group of people so excited and passionate about their jobs as the people in CodeGear. We're relishing the chance to change the lives of software developers, and maybe have a bit of fun along the way. There has definitely been a return of the old "Borland Barbarian" spirit within the CodeGear ranks. However, the proof will be in what we deliver, so judge us on that, not on rumours and comments from other places.