I bought an OLPC on their BOGO program. I noticed the touchpad would start to act up after a short time using it, but for the price I could afford to overlook it. Then I started having trouble connecting to the wireless. Soon it wouldn't connect at all. Every so often I'd spend a few hours troubleshooting and updating the OS, to no avail. Eventually I got the developers key and ran a hardware check. The hardware check would lock up every time at the same place. The OLPC wouldn't connect wirelessly. I now wonder whatever happened to the second laptop I bought for the third world. Did it have the same hardware failings as mine had? Likely. And why is it now, several years after the OLPC was first offered, that I'm finding out the problems I experienced were built into the OLPC through a series of incompetent design and engineering compromises? People are coming out of the woodworks now, but where were they then?

SQL isn't really a programming language. A programming language is one which is either interpreted or compiled, and is capable of manipulating data across a variety of problem domains. SQL is useful in only one problem domain---manipulating data in a relational database management system. In reality, SQL doesn't even do that. Instead, it is a tool that tells the RDBMS how you want the data manipulated.

The initial thread discussed the requirement to use only one language. In practice, this is impossible. Web development is illustrative. You can't build a modern website in only one type of language---you need HTML and/or XHTML, you need CSS, you need Javascript, etc. If you proscribe the use of those tools, you limit the potential of your web development.

In like manner, to prescribe the use of one particular language for all development puts artificial limits on the company's potential. Java has its uses, but doing game developement or scientific programming in Java is a waste of time and resources. For some high end programming, where compute cycles are at a premium, you want to use a language like C. For most programming, a language with automated memory management and garbage collection is best, because compute cycles are cheap, but programmers are not.

I wish I knew which companies were forcing the use of any single language for all development efforts. I'd also bet that their development costs would rise as their programmers struggled to use that one language for tasks it was ill-suited for. I'd wonder if their increased costs would push the company into moving programming to the developing world. I'd wonder if they were able to attract and keep the best programmers, who would likely despise working in that environment. I'd wonder what would eventually happen to their stock price and, eventually, the CIO.