[sc34wg3] Re: Whitespace agnosticism in CTM and descendants

[Lars Heuer]

Hi Xuân,

[...]
> This problem is resolved in LTM in making a difference between : on the
> one hand and :  respectively  : on the other hand. The differences

Right, the same was made in CTM and it was discussed in the Montreal
and Leipzig (2006) meeting.

In Montreal the committee came up with a solution to use "bangers" for
QNames, i.e. foo!bar. But in Leipzig the committee decided that the
colon overloading is not such a big problem and we should switch to
the common notation for QNames, which is foo:bar.

Regarding the roles: We know that the colon delimiter is not strictly
necessary to delimit the role type and player. In fact, we
experimented with a notation where we the colon was not used.

To make the story short: The committee decided that an average
simple-minded user (a.k.a. the typical Topic Maps user ;)) should be
able to recognize the difference and that we use the common notation
for QNames and we leave the colon as type/player delimiter for
*readablity*.

[...]
>    5. Overload the colon and make the life of CTM authors and CTM parser
>       authors unnecessarily harder, more troublesome and error-prone,
>       forever.

I'd go with 5. ;)

I don't think that the we are making the life of CTM authors harder.
The parser-writer has to make some (not very difficult) decisions
regarding QName and role-type/player combinations, but IMO it is
acceptable.

[...]
> If we do not fix this bug soon, everyone will get used to the bug (maybe
> some of you already got used...), and the bug will become permanent,
> plagueing every CTM, TMCL, TMQL author for ever. So, please, let's fix

Topic Maps apocalypse ;)

Seriously, I think it is a non-issue. The EBNF is currently not valid,
true. But the next draft will fix it. If the committee brings up this
colon-issue again, the editors have to fix it, but the editors do not
plan to change the current colon-usage.

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