Narrowing field of languages?

Looking at the posting histories for the 58 languages listed on Google Groups, I see two notable trends. First, most languages show a peak in activity around 1999-2001. I've always assumed this was because technical users were moving to subcription-only groups to avoid the increasing noise on Usenet.

The other trend is a dramatic fall-off of activity in several established languages in just the past 2-3 years, particularly Visual Basic, Eiffel, Java, ML, Pascal/Delphi, POP-11, Rexx, and Smalltalk.

The older languages that have been notably holding steady are AWK, C, Forth, Fortran, Prolog, Scheme, Verilog, VHDL. Lisp is a special case in that it was dying in the 1990s then revived around 2000.

It seems that languages dominating a special niche are keeping their users while general-purpose languages are suddenly giving way to a few new superstars, particularly Python and Ruby.

This would indicate that the future of general-purpose programming is to be marked by much less variety in languages than we had in the past. But I notice that functional languages like OCaml and Scala are generally not represented there. So perhaps this sample is misleading?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Just a hunch...

I suspect the sample is misleading, although I won't speculate as to why. In general, prophesies of "convergence" always seem to come on the cusp of diversification (I won't say balkanization).


There are no groups for major languages like C#, F#, OCaml, Groovy, Scala, Erlang, and Lua. It's likely that new communities are simply using other than Usenet. I just wanted others' sense of whether something more significant might be going on. It is striking, though, the apparent sudden shift from old favorites in just the last 2-3 years.

Java and VB..Java isn't so

Java and VB..Java isn't so dead yet, its just "mature" and no longer the hot new language to learn as everyone already knows it. Some of the languages you mentioned are hyped but not widely used (Erlang). Other languages have simply been dying for a long time and are finally being forgotten. Maybe you are following the hype too closely (I'm assuming activity == hype), instead the reality is actually quite boring (unfortunately).

Less posts since...

Questions and answers are now indexed by Google, so there is less need to post?


For what it's worth, it seems to me that traffic on the Scala mailing lists has grown a lot over the past year or so. Your picture might be more complete if you figure out ways to add some data mined from gmane, nabble, yahoo groups, and whatever are the other popular email list providers and aggregators.

targeted search trends

Perhaps compare web search trends for error messages from different languages and popular libraries for them... book sales, news articles, forum discussions, open source repos, job postings, etc. aren't good gauges for the use (or lack thereof) of a language.

For example, I probably won't buy a book on JS at this point, might write about a language I'm making, hypothesize about others, make an open source project to learn another, post duplicate job requests because it's hard to fill a niche language position, etc. However, if I'm writing production code, I'm googling on compatibility, implementation details, error messages, APIs, etc. Many studies cite search trends of language names, which might make sense (hits "ocaml unmatched case.."), but this seems noisey.

Can you provide your data?

Uh, please?