To make this site work we need very specific questions. Requirements that can be measured and answered.
Can there be a question with a degree of too specific requirements - meaning requirements that maybe nobody ever has again?
In other words - Do I need to be careful how specific I am while asking?
As long as it's an actual problem or decision that you face, you're fine. This site is going to get questions that are just hard to answer. This could be for a number of reasons:
The whole point of 'Too Localized' was to get questions that people might trip over cleared out of the way. These questions might be great matches to words that you typed in a search engine, only to find that your answer isn't there - it was just a (missing semicolon, bug in software that was fixed ages ago, otherwise completely unrelated issue).
Where this gets interesting is bugs, but this isn't as much about using software as it is describing what you hope the software would do for you. Even there, I don't think we have to worry too much about the premise of your question actually ending up to be out of the scope of the question itself.
On this site, I don't see it becoming a major problem.
As stated in the comments, just being "too localized" is no longer a valid reason for closing on Stack Exchange (a fact that I believe to be a good thing--how would you feel if you came here for help only to be shut down because no one though your problem would be relevant to other people?)
Additionally, we've consistently emphasized precision and specification when it comes to asking questions here. Without these policies, the whole site would quickly fall apart into an endless slew of "What is the best X?" questions.
Ultimately I think it's better for everyone if just forget the idea of considering a question being "too localized." The only alternative is to start telling users: "make your questions precise, but not too precise," which is in itself an imprecise, subjective judgment call.
No.
There is no requirements that are too localized.
There is just software that does not exist yet.
If requirements are extremely hard to satisfy (for instance: Pi generator that takes less than 1 second per million digits and takes less than 640K in memory), no answer will appear, until maybe some day when someone writes a tool that does this.
