What is SDO?
Shortly after the September 2004 Borland Conference, I caught up with Borland's Dale Fuller and Nigel Brown. In talking to them, I was particularly interested in following up recent announcements about SDO (Software Delivery Optimization), a new Borland initiative for which the company has high hopes. The first the world got to know about SDO was in a blog by Danny Thorpe, Chief Scientist in the Developer Tools Group, who wrote on August 16th 2004:
"How do you measure a market that has been hiding in plain view, that's on nobody's radar? … It's big. Big work. Big risk. Big reward. Caution: Disruptive ideas ahead."
SDO was unveiled on September 13th at the Borland Conference in San Jose. I believe it’s fair to say that developers were underwhelmed. In fact, for some the announcement went unnoticed and afterwards they were still asking Danny what the big project was. The problem is that the SDO concept is so full of corporate buzzwords that it tends to be ignored by real-world developers, who look for more tangible things like new product features. I therefore took the opportunity to press these Borland execs on what SDO really delivers, though product details remain vague.
Dale and Nigel do talk about a forthcoming product called Themis, due in the first half of 2005. However, it's likely that most of what is in Themis is not new, but the company’s existing ALM (Application Lifecycle Management) products repackaged. Judging by Boz Elloy’s remarks here, we will have to wait for the follow-on Project Hyperion - “visibility, predictability and decision support” - and Project Prometheus - “ERP for software delivery” - before we know what SDO really amounts to. The discussion hinges on one radical question. Can you treat software development in the same way as other design and manufacturing processes?
If you don’t care about SDO, you’ll be glad to know that we touched on other topics too, including 64-bit development and targeting mobile devices. The menu on the left has links to the various sections.
Tim: The Software Delivery Optimization white paper says Borland will “facilitate a heterogeneous Application Lifecycle Management environment”. But how will you implement that?
Dale: ALM is an environment for team-based productivity. SDO is the bridge that crosses the chasm between business and the technology organizations within corporations. Latest studies show that 80% of all IT projects are classified as failures. That’s gigantic. So in our research we started diving into it and looking at why the failure rate. We found the single biggest point was a lack of communication and collaboration between business and technologists. So with SDO what we wanted to create the bridge. Our goal is if we can just cut the failure rate by 10%, from 80% to 70%, that’s a gigantic uplift in productivity.
Tim: Can you further clarify the difference between ALM and SDO?
Dale: ALM is all about the product, the technology. SDO is about the interface of business to the technologists. That's why we don't want to get people confused and say, "Hey Mr CEO, we’re now going to have ALM for you." He's going to say, "I'm not a product manager, I’m not going to create source code." As the CEO what I want to look at is a dashboard. So SDO is much more of a monitoring, guidance system, from top management down into the organization, and for the technologists to look back into business and learn "Hey, this feature we were going to cut out, it actually has the highest business value. So we should focus on that first."
Tim: So how do you make sure your development team is reporting correctly to its supervisors? How do you make sure that the requirements are really coming through to the development team? How do you get the feedback from the customers to become a change request? They are all difficult management issues. Has Borland really got a silver bullet that’s going to address all these well-known problems? Will you provide software, or services? or methodology?
Dale: It’s a little bit of all those things. One of the things that I’ve learned in my life is that it’s never one silver bullet that solves all problems. It’s a lot of incremental stuff. Typically it’s the easiest steps that people overlook. An example is an ALM strategy. I wish I could say that it is our strategy. But Rational came up with this strategy seven years ago. They never implemented it. Two years ago we took that strategy, put it together with our products and integrated it all together. For example, now my requirements are written directly into the code, I can look at the code and make changes to it and it updates the requirements. Nowhere else in the world is this done. I can make changes to the model in real time, and the code starts changing and the requirements start changing.
What we’ve learned is that communication and collaboration are the simplest things to do, if you have the technologies that allow you to do it. SDO now extends that into business. What if your software architect makes a change to a feature set that actually changes the entire value of a project you’re working on? It happens because technology people make decisions that a marketing person should make. When it gets shipped, it's the wrong thing.
Tim: I understand the problem you're describing, and it resonates as being a real problem, but what I'm not clear about is how Borland is going to address it.
Nigel: One of the things is that we want to get the software to do the checks and balances rather than this human interaction which is prone to failure. That is why we're deeply linking all the products. That's where we're unique, because nobody else is doing the deep integration.
Introducing Themis: Click here for part 2 of this interview
Copyright Tim Anderson October 15th 2004