> the purpose of that programming is directly to solve a computer science problem.

This is not quite it. The asker may be trying to solve a problem in their domain using computer science (tools). In my opinion, tool support questions are feasible here if

 1. the question requires CS expertise (that is not only tool expertise!) and
 2. the tool the asker uses is an artefact of (contemporary) computer science.

(1) can (but does not have to) imply that a good answer to the question should carry over to other, similar tools. For example, a question asking how to do something in proof assisstant X can cause answers that can be applied to proof assisstant Y. We have had examples of such questions, iirc.

Regarding (2), proof assisstants and SAT solvers are, arguably, artefacts of computer science. Most other tools are not. Compilers can be seen as such; here it depends on which aspect the question touches. Questions about the principles underlying a certain optimisation in the gcc are ok, questions of the kind "What does compiler flag -g do?" not.

---
I don't know anything about the tools referenced in [the linked question](http://cs.stackexchange.com/questions/2510/bit-blasting-from-smt-lib2-format) and the OP does not give much information, but it seems to me that (1) is violated.