Thread from comp.lang.tcl sharing one person's positive experiences trying out Python after getting frustrated with the generally poor/abandoned state of many Tcl add-ons (something I'm getting bit by a little at the moment). It's somewhat inevitable given the lack of coolness; inevitably you end up losing the younger developers, always wanting to jump on the next new thing. While the core language is wonderfully mature and in great shape, the ecosystem as a whole is certainly not. (Note that I also think that more and more people and companies are doing stuff internally, and not releasing the results back... which is okay, but on a large scale not healthy for the ecosystem).
This continual language/tool churn in the industry isn't good for anyone. People are too quick to jump to a new thing for what really amounts to the promise of a 10% net improvement, but it takes years of pain to really get anywhere near even that modest goal, factoring in all the non-glamorous "mature" stuff that was already done in the previous language/tool, none of which can be easily moved to the new environment. From Tcl's case, think I18N, a solid I/O model, true cross platform API's, a ton of performance work, and deployment, just to name a few.
Addendum: Good posting addressing underlying causes.
Comments