In response to the July 29 Perspectives column by Jon Oltsik, "":
You make some good points. However, the problem goes much deeper. People forget that legacy systems have taken as much as 30 years to build and those resources are now paid for and gone. To put something new in place would require massive costs (as well as risk failure). There are a couple of linked issues to consider:
One is know-how. Thirty years ago, people from the industry were writing the software, using easy-to-use 4GL languages. They had the knowledge to know how the internals of the business worked, to deliver robust applications for users.
The other is technology. Show me a C++ developer (or .Net or object-oriented guru) who knows how to create a full round-trip business transaction (boundaries) across the same system or multiple platforms. They just don't have the mindset to learn the business and go talk to users--let alone implement something that other people want.
These are the biggest barriers to new applications and make management fearful about allowing IT loose on something--however archaic--that works. I am on the prowl for the tool that will allow 'real people' to build applications rapidly, just like we did 30 years ago.