Is it just that the language is too far in language-space from what you are familiar with? That there are too many new paradigms involved for comfort? That may be why the ability to “write COBOL” in your new language is such a soother. It lets you play in the language and learn the new idioms as you do so.

I still don’t have a decent replacement for “too academic” though :-(

One other thought (which could help prove or disprove the above) is that it would be interesting to hear from someone who learned programming with a functional language about their experiences moving to Java for the first time (“What do you mean, I have to tell it what to do?!”) I suspect that the dificulty would be vastly reduced because that form of imperative instruction is how we tend to deal with each other. Instructions stating how to make a cup of tea are much more likely to look like C than Haskell (well, maybe not C, but you get the point ;-) )

In saying we should be sensitive to the limits of gradual adoption, I’m talking about building hand-holding into language infrastructure based on the assumption that someone knows a mainstream language but is learning yours. Useful compiler/runtime error messages, an accessible API, and, yes, even some language features to mimic dominant paradigms[1] are all ways to do that.

[1] See Brian’s OCaml example above — I couldn’t remember the for loop syntax when I wrote my comment.

I’ve re-read your post a couple times. It’s interesting to hear that Groovy has been more successful than Scala at being adopted, and I hear you to account for the difference by saying that Groovy has facilitated a more gradual adoption route.

Yes, I’m partial to the “Intuitive Means Familiar” idea. It’s not that I think people who fend off so-called “academic” languages are lazy, though. I think each person works from their own cost-benefit analysis, and (like Brian) I think that the estimates of high difficulty tend to be what put people off.

I agree that it would be better if “academic” were not the term used as shorthand. It’s inaccurate and counterproductive.

I also agree that when a language “pushes the envelope” it should not be a surprise when it does not sweep the industry as the new big thing.

What I remain dubious about is the ability to “be sensitive” about the limits to gradual adoption. Are you talking about actual changes to languages, or are you talking about improvements to identifying and communicating ways to help build skill in the (unaltered) languages?