[sc34wg3] CTM for non-SVO languages

[Lars Heuer]

Hi Kevin,

[...]
> My issue was that I was told that in some cases the only way to end a
> template was with a double line-feed.  I said that counting on line-feeds as
> terminators was bad practice from a quality management perspective.

The line-feed in templates were never a problem, since we have the
"def" and "end" keywords to decide where a template starts and where
it ends.
If you refer to the "end of topic" marker, you're right:
We have had two possibilities to mark the end of a topic:
   a) the dot (.)
   b) an empty line

The empty line was seen as problem by several parties, so we added it
to the CTM issue list and asked the WG, if we should just allow the
dot to mark the end of a topic declaration. I think, the "." works
very well and is used by some other syntaxes as well.

> So, when curlies were proposed, I felt that curlies solved that
> issue. If curlies have been cancelled, then I want to make sure that
> we have solved this issue that I raised about using line-feeds as
> terminators.

  john {
  - "John Lennnon" }
  paul {
  - "Paul McCartney" }

is equivalent to:
  
  john
  - "John Lennon" .   # End of topic "john"
  paul
  - "Paul McCartney" . # End of topic "paul", no whitespaces
                       # necessary.

                       
> The more recent discussion about non-SVO languages has me suspicious that we
> are relying on line feeds as separators as well.  I can't tell for sure from
> the discussion.  But, if we are, this is an equally bad practice from a
> quality management point of view.

This isn't the case. Meaningful EOL markers where only used as
end-of-topic marker and it's likely that this issue will be solved
either by the dot or by any other solution (curlies etc). In practise
it makes sense to use whitespaces and EOLs, but CTM does not enforce
EOLs (technically).

(We've one more meaningful EOL problem in the current draft (the
mergemap directive), but this is easy to solve and is not a big
problem at all.)

> So, if those who oppose curlies and semis could address these issues in a
> post, I would be most grateful.

I hope this post clarifies the issues a bit :)

Best regards,
Lars
-- 
http://www.semagia.com