FW: [sc34wg3] CTM: Working draft for Montreal

[Lars Heuer]

Hi Andreas,

[...]
> - It might be worth of allowing the "empty assertion", since experience
> (e.g. with CSS) has shown that using, e.g., a semicolon as separator
> instead of as terminator can be confusing to authors; it would be all to
> easy to forget changing the period to a semicolon when appending
> assertions, for example.

Some of the editors are aware of the problem with the semicolons and
have proposed an alternative syntax (with no semicolons) that utilizes
the colon as type/value delimiter.

     subject
         some-occ-type: "Occurrence Value"
         another-occ:   http://www.example.org/

Note that "colon" means here: Colon + min. one whitespace (':' WS+).

The proposal was rejected because QNames are using the colon as
delimiter, too.

    # Example using QNames
    subject
        q:name: "Occ Value"
        q:name2: http://www.example.org/


    # Example using QNames with optional whitespaces to
    # enhance readablity
    subject
        q:name  : "Occ Value"
        q:name2 : http://www.example.org/


The colon 'overloading' was considered as bad practice, difficult to
teach etc.

<personal-note>
IMO it is easier to remember a delimiter *inside* a statement than at
the end of an statement. If I type something and consider it as
"complete" my mind jumps to the next issue without taking care that
the previous statement is not complete because of a missing
end-delimiter.

I think a missing end-delimiter is one of the prominent failures
newbies make while learning programming languages.
</personal-note>

Anyway, if someone has a good alternative syntax that does not require
(end-)delimiters, feel free to post a proposal.

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