An article in last month’s Infoworld strikes a somewhat negative note on Microsoft’s Software Factories initiative. Despite hitting 100,000 downloads since release 6 months ago Greg DeMichillie, lead analyst for software development tools at analysts Directions on Microsoft, is sceptical.

DeMichillie compares Software Factories to technologies such as DLLs and OOP and notes that those technologies only offer incremental improvement. He doesn’t see Software Factories as offering the step change that the industry needs.

However, DeMichillie seems to have failed to pick up on one of the key differences between Software Factories and earlier reuse-based approaches. Software Factories, and other Software Product Line approaches, derive their benefits from a strategic approach to reuse based on economies of scope.

What this means is that applications derived from a Software Factory are all sufficiently similar to allow reusable components to be developed once and incorporated into many applications. This economy of scope means that the cost to develop and reuse these components is lower than if the applications were unrelated.

DeMichillie also seems to be unaware of case studies from the community that show order of magnitude improvements in key measures such as productivity, quality and time-to-market when organisations transition to a Software Product Line approach.

Whether Microsoft’s Software Factories initiative will be successful or not is another matter. In pure::variants we have our own approach to building and managing Software Product Lines which takes a slightly different direction from Software Factories - although in fact pure::variants can complement Software Factories and other DSL toolsets such as Metacase’s Metaedit+.