Borland goes for ALM 2.0
Is it deja vu all over again?
Nine months ago, Borland announced it was spinning off its IDE tools into a separate business to focus on ALM (Application Lifecycle Management) products and services.
The divestment is taking longer than planned and, as I write, the Borland Developer Tools Group is still part of Borland. But is the focus on ALM still on track?
Recent reports that the company is abandoning its Core SDP (Software Delivery Platform) suggest some uncertainty; but Marc Brown, director of product marketing - ALM Products, assured me this is not so. The changes are apparently more to do with licensing and packaging than with the strategy itself.
Core SDP is a role-based product with different features available according to whether you are an analyst, architect, developer, or tester.
Borland discovered that role-based products do not work. "We found that the fixed role-based packaging that we offered in Core SDP simply was not satisfying customer needs. Customers define those roles differently between each organisation or in some cases even between different project teams," Brown said. Perhaps Microsoft will learn this same lesson: it has fallen into the same trap with the licensing of its Team System suite.
Still, this does not change Borland's central theme of SDO (Software Delivery Optimisation). SDO is a phrase which hard-core developers tend to dismiss as being a buzz word too far. If so, they will also dislike Borland's new meme, ALM 2.0.
In truth, SDO and ALM 2.0 are nearly synonymous. Forrester Research appears to have coined the term ALM 2.0, in a paper on the coordination of development lifecycle activities here. It is about integrating the outputs from multiple tools into a single, manageable, and predictable business process. This is similar to what former Borland CEO Dale Fuller described to me (see here), back in 2004 when SDO was first announced: a dashboard from which all project participants could monitor and analyse progress.
Brown said the dashboard concept is "still a huge theme". "Organisations lack the ability to have objective measures that can be surfaced within a dashboard, to give them the objective data to manage these projects correctly."
Getting such a dashboard to work is a significant technical problem, particularly when organisations use ALM tools from a variety of vendors, ranging from modelling through to requirements definition, development, testing, profiling and change management. It implies a standard API through which such products can report their status and interact with other tools, and no such API exists.
Microsoft's Team System, with its project portal and reporting tools is an interesting single-vendor approach, while the Serena-led ALF (Application Lifecycle Framework) project at Eclipse is working on a common vocabulary and eventing system for lifecycle tools using web services.
Despite its love affair with Eclipse, Borland is dismissive of ALF. "Adoption has not been positive by the bulk of the ALM vendors," Brown told me. Part of the reason is that Borland is developing its own alternative. "We've developed an ALM metamodel that we'll use," Brown said, including its own "vocabulary to share data across the application lifecycle".
Borland intends to work with ALM partners such as Mercury to build adapters into a federation layer, or else to build its own adapters where necessary.
It is an ambitious plan, and makes you wonder why Borland does not put its effort into ALF to make broad adoption more likely. However, it remains a promising approach, even if in the end it works best with Borland's own growing range of ALM tools.
I also asked how Borland is untangling its ALM tools from the IDE products which will eventually belong to another company. "It was a challenge to isolate what was being shared, what was needed by both sides, who the owner should be, and what agreements needed to be in place," he admitted, but added: "It was actually healthy. It was a good clean-up exercise."
The spun-off organisation will, it is planned, remain a strategic partner. ®