@Tim: Yes, a technology can succeed by diffusion and fliw-on effects I guess. Winning by losing? 

When a programming language gets more "markup-y" with annotations allowing declarative multipurposing, position-independant named fields, unicode and multiline strings, it increasingly becomes useful for some things XML is good at.  

I think I have mentioned before that (going back 30 years, when Java was still Oak!) when I worked at TI supporting their LISP hardware (we never sold any in Australia, so "support" is a bit of a euphemism) a US LISP guru said to me that LISP would win by losing: he said, entirely correctly, that the future would be all "LISP is dead"  C family syntax languages which, if you look under the hood, would be LISPish: garbage collection, objects, eval(),  and standard annotatible lists (e.g., DOM trees).  Truest prediction I ever heard.

But now I think the reverse is also true: a language may "win" by losing and become moribund, but be resuscitated by itself taking advantage of the other ideas floating around. (I guess Clojure is a recent example of this, for LISP.  Indeed, _javascript_ was languishing and being outpaced 20 years ago, until it started learning from its peers and distant cousins. Never say never?) 

Rick