[sc34wg3] Semicolon or not semicolon

[Dmitry]

On 29-Jan-08, at 2:00 PM, Lars Heuer wrote:

> Hi Dmitry,
>
> [...]
>> I would prefer it this way:
>
>> paul
> [...]
>>      last_name - "McCartney";
>>      isa person, musician;
>>      o:works-for    [o:The-Beatles - "The Beatles], o:The-Wings;
>>      o:homepage  <http://www...>
>> .
>
> Yeah. I don't want to reject your input, but IIRC we were at this
> point already as we discussed the Montréal proposal to use colons
> everywhere. The result of that hard and long discussion was:
>
>   - isa/ako are keywords and do not require colons or parenthesis
>   - Template invocations require parenthesis
>   - Occurrence types / name types require a colon
>
> These resolutions were the starting point for N0935.
>
>
> To your proposal:
> - QNames for templates do not work, since Montréal and Kyoto proposed
>   to remove the template-import-directive in favour of a revised
>   include-directive
>

"o:work_ for" is not a template in my case, it is a binary  
association with standard interpretation

o:work_for(tm:subject  paul, tm:object: TheBeatles)

Inverse relationships can be represented using "^" prefix

So

TheBeatles
    ^o:work_for   paul.

is mapped to the same thing

o:work_for(tm:subject  paul, tm:object: TheBeatles)

Another option is more symmetric

TheBeatles
    o:work_for   (paul,_).

Again, it is not a template, it is a "real" association with  
"standard" roles.


Symmetric  associations can be handled with additional prefix "&"

john
      f:friends &paul.

will be mapped to

f:friends(tm:subject :  paul, tm:subject : john)


This proposal will cover probably 80% of associations without templates.
It  will also simplify RDF interoperability.


> - A comma does not work for a list of values, otherwise the following
>   is undecidable:
>
>     occ <http://a.com/> @<http://b.com/>, <http://c.com/>;


One of the suggestions was to use "()" around complex scopes

occ <http://a.com/> @(<http://b.com/>, <http://c.com/>), <http://d.com>;

>
>   How many occurrences do you count here? Belongs <http://c.com/> to
>   the scope of the occurrence or is it another occurrence value?
>
> I think N0935 and the Kyoto proposals should be the starting point for
> a discussion and not a completely different syntax. Sure, we can
> discuss a completely different syntax but we hold up TMQL and TMCL and
> we're running out of time, IMO.

I think we have a good chance to make CTM a little bit better than it  
is now. It will be more difficult
to change it later.

Dmitry