Re: [Opencxx-users] Re: refining the Synopsis APIs
Stefan Seefeld <[email protected]>
| Newsgroups | gmane.comp.documentation.synopsis |
|---|---|
| Message-ID | <[email protected]> |
Vladimir Prus wrote: >>I do believe that the names should express that one is focussing on syntax, >>the other on semantics (though the meaning of these two terms is definitely >>more clear if used in conjunction, i.e. opposition, and not independently). > > > Excellent! What about naming the modules "Syntax" and "Semantic"? The advantage of 'AST' over 'Syntax' is that it is well recognized in the community / literature. That's not quite true for 'ASG' (even though there are quite a number of places using the term, in particular in conjunction with 'AST'). I find 'Semantic' not very satisfactory, not least because it is not a noun. Semantic what ? Isn't our struggle also related to the fact (as you once stated yourself) that there are quite a lot of different types of semantic information to be extracted from the sources ? May be we get further looking for more precise terms for what we try to express there. There are other kinds of data to be extracted that would equally well fit under the 'semantic' umbrella, such as call graphs, control flow graphs, etc. (and may be someone will want to build these representations on top of our current work). How would you describe the representation currently used in the Synopsis.AST and Synopsis.Type modules (i.e. the data structure used to generate documentation) ? Regards, Stefan