Re: [Ctags] Same extension ?

[Darren H. <da...@da...>]

On Sep 29, 2007, at 11:13 AM, Elliott Hughes wrote:

> a bigger problem for editors, in my opinion, is that every language  
> gets
> to choose its own kind letters, so to know what a tag is, you need to
> remap all the kinds to an internal set of kinds.

I agree with the spirit of this suggestion, as it is inline with my  
general philosophy. However, from my perspective, there is not enough  
commonality between unrelated languages to have a universal set. In  
fact, I might be willing to wager that you cannot find even a single  
kind that is common amongst all languages (the closest candidate  
would probably be "routine").

Now, when I say this, let me clarify that in trying to come up with  
something that is common amongst all languages, one would be forced  
to talk about abstractions of concepts that might have specific  
implementations in each language. But what that would likely mean is  
that they would no longer be named for the concept familiar in the  
nomenclature of each language.

For example, C calls every routine a "function", but this term has  
specific meaning in languages like Fortran or Eiffel, where not every  
routine is a function.

Right now, the kinds are named according to the nomenclature familiar  
with that language. If we come up with an abstract set that is the  
superset (i.e. "mother") of all kinds, they will be named in ways  
that might be foreign to users of some languages (at best), or  
contradictions with the established nomenclature of that language (at  
worst).

Perhaps you might try your hand at generating this abstract set as a  
straw man. Or clarify how I have misunderstood or exaggerated the  
challenge.

Darren

--
Darren Hiebert
http://darrenhiebert.com