I have finally installed a version of CA-Superproject under W98/Virtualbox and the experience reminded me of a couple of things, about the software, about its final custodian, Computer Associates (CA) and also some critical software project management issues. I have written a more formal note on Linkedin and this is my mirror/pointer to that; the rest of this article précises that article. For more, see overleaf/below. …
I like it because it was good with both time and money, it had a PERT/CPM interface and thus was good with task dependencies, it was good with cash flow, had some Earned Value Analysis tools and it had some powerful calendaring tools.
Its death is partly as a result of the victory of the integrated suites and while CA may have assembled the components to offer such a product, they lacked the engineering capability to integrate the products into a single compelling offering and both the technical marketing capability and marketing will & power to sell into the desktop market. They also lacked an OS and thus were locked out of the OEM market.
They were not the last to fail to understand that the economics of software requires the volume of the desktop, and today the pocket. In the end, the desktop market was too hard to compete in, for a company organised to sell high ticket software bundles to business. This article in the NYT shows how CA was captured and trapped in a licence sales model.
Superproject was designed at time when Labour was fungible, this is less true in the software engineering business and assembling a team of three engineers where often/usually the individuals assigned cannot substitute for each other is a more difficult scheduling problem. Managing this problem is hard and to my knowledge no project management software does this well.
One of the reasons that the more simple MS Project took over was, I believe, the poor project management techniques used in information systems project engineering although Microsoft’s ownership of the mindshare of the project management community cannot be discounted. By poor techniques, there was very little understanding of how to place a value on interim deliverables in software projects and performing “earned value analysis” was hard or impossible. The limited fungability of Labour was another problem that neither product solved. When we consider that time i.e effort is money and that much work cannot be speeded up by assigning more people to it, we had new problems to solve and needed new tools. The truth is that neither MS Project nor Superproject answered these questions.
Another product, not good enough to survive which ended up in the graveyard curator’s possession.