From Dan Mabbutt,
Your Guide to Visual Basic.
FREE Newsletter. Sign Up Now!

Dec 10 2006

Visual Basic Plummets as Java Takes Lead among Software Developers
--Tekrati.com, December 6

Visual Basic Use Eroding
--CMSWire, December 5

Visual Basic Usage Down
--ars technica, November 30

The headlines fairly scream that the long reign of Visual Basic at the top of the heap of developer technologies is over. Here's a typical quote: "Microsoft has dominated languages since the early 90's but we are seeing much more parity now."

Take that, smarty RAD language! Nyaaa, nyaaa, na-nyaaa nyaaa!

All these headlines have a single source: a new survey from Evans Data Corporation, a market research company. Based on returns from 430 (wow!) developers, they have concluded that developer use of the Visual Basic family has dropped off by 35% since last spring. VB.NET is doing slightly better. They're only down by 26%.

According to Evans Data, who is at the top of the heap? Java is the big winner. They say 45% of developers use Java during some portion of their time. C/C++ is at 40%, and C# is at 32%.

Evans Data has this to say about the trend for VB.NET in particular:

-----------------

Use of VB.NET had not changed in the last three reporting periods. Adoption had stabilized at 30% with less than 1% variance. During the current period, the use of VB.NET has declined more than 7%. The decline should come as little surprise given declines in related languages and tools, despite continuing predictions that VB.NET would serve as a preferred migration path for these and other end-of-life languages.

Also ...

The use of scripting languages ... appears to have limited VB's future market potential.

-----------------

My dad used to say that, "There are lies, damn lies, and statistics." I won't go that far, partly because I have devoted a lot of hours of my own education to studying statistics. But there's still a kernel of truth to what my dad said.

For example, the statistics screaming from most headlines - percentage of decrease - are statistical "funny numbers" in this context. As as example, consider that a drop from 2 to 1 is a 100% decrease! The number that should be quoted is absolute market share which, according to Evans Data, only dropped 7%.

In addition, you have to read their detailed methodology page to discover that the formal mathematical margin of error, plus or minus 5%, isn't that far from the actual measured result, 7%.

The unavoidable problem that is more serious is that these formal mathematical results are built on the sandy foundation of human responses. Statistics are most reliable when they measure something hard and physical: the changes in the radiation of the sun; the movement of continental plates; the growth rate of bacterial cultures. As election pollsters know only too well, they are least reliable when they measure human responses gained through an interview process.

Still, there's "probably" (there's that statistical point of view again) something to these results. What could it be?