Why do I raise these issues now, if they're not new? Two reasons: we're in the middle of a massive economic slump, and making poor software-buying decisions is costing us money that could be spent on something that contributes to growth (or even survival!); and secondly, there have been some more recent technology changes that make old habits of software solution selection anachronistic.
The core of the problem that we need to unpick and then fix is the totally misguided belief that the best software is written by multinational organisations. We have the ludicrous situation in which no one is sacked for buying SAP. Well done, Mr IT director, you chose the global best-practice software. Now all we have to do is change our entire way of working and all of our business processes.
This is not to attack SAP - the company makes an extremely fine product - but far too often the reason to choose one of the big box software vendors is just craven - and often plain wrong.
The idea that there is a single best practice or process model is obviously a fallacy. Each company is different and each country is different, very different.
Yet the stories of massive tenders for large-scape enterprise systems being won by the usual global players continue, despite the non-stop litany of cost overruns, underperformance, and non-stop goalpost shifting.
This is not new, so what is the fulcrum that we can use to start to shift industry choices? Where is Archimedes' "long enough lever" that we can use to change a culture of enterprises opting for the safe choices? The safe choices that doom them again and again to wasting precious resources, and dooms us in the local software industry to a bit part in the great drama of enterprise solutions.
The lever that could move the software-buying world is a rapidly maturing service-oriented architecture environment and increasingly popular cloud computing models, along with modern end-to-end system-monitoring tools.
There used to be one particularly strong argument for a single-vendor mega-solution from an SAP, Oracle, Microsoft or IBM. With a single vendor you had a reasonable chance of being able to manage your whole infrastructure and that all elements would reliably talk to others (good luck with that, in the real world).
Now, however, with SOA and cloud technology we have a model that allows us to build best-of-breed systems that are seamless, reliable and easy to troubleshoot. We can buy point solutions where there is a good fit with the business' existing processes and bring in custom development from a strong local software development house when no shrink-wrapped product can be found.
Where there are conditions peculiar to South Africa we can easily engineer around them: we can develop around unreliable power, or limited data centre capacity, or limited connection speeds.
So often an enterprise evaluation team will look at the major providers of software and rate the fit. They'll use the old rule of thumb that if something is less than an 80 percent fit, then it's better to start from scratch. And, guess what - they go overseas on fact-finding mission and they come back home having found an 80 percent fit. Except when they implement the system it turns out that it was actually a 45 percent fit and millions are then spent to re-engineer the business to make it work.
Of course, no one will be fired. So nothing changes, but it's time it did.
We've got some really good and exceptionally innovative software developers in South Africa, with a culture of "get the job done" that goes back to sanctions days when they had no choice - and that has been tempered by the last 10 years of having to compete against global players entering our market.
The bottom line for my business, which provides enterprise software -development tools to South African development houses, is that our main competition is brand bigotry.
Every day we have to work like demons just to overcome the fallacious assumptions that global software companies make better technology than local ones before we can even start a useful conversation about what the enterprise needs to achieve its goals.
It's time we took this issue head on - it helps no one. It's time we followed the example of our peers in the hardware and networking world - 21st-century technology means that we can implement the systems that work best, we are no longer taking a massive risk buying best-of-breed in software.
