[sc34wg3] CTM: Semicolons

Lars Marius Garshol larsga at garshol.priv.no
Tue Nov 4 08:32:25 EST 2008

* Jaeho Lee
> I think there were other technical reasons mentioned in Oslo more  
> than just readability for non-SVO languages.
> One of them is easiness for the parser, and more importantly  
> easiness for
> the generator of CTM statements.
> My many application programs will generate CTM statements routinely,  
> and
> handling optional component is always headache.
> Furthermore readability and ugliness are so subjective issues. I see  
> the
> beauty of consistent when there are always semicolons at the end of  
> every
> statement.

It's certainly true that these are legitimate technical arguments for  
requiring the semicolons. It's also true that whether the semicolons  
are beautiful or not highly subjective.

For me, the strongest argument in favour of them (inside topic blocks,  
that is) is that it's likely to lead to better error messages when  
there are syntax errors in bigger topic blocks. (Note the "likely". I  
haven't tried this out in practice.)

--Lars M.

More information about the sc34wg3 mailing list