On Thu, Mar 04, 2004 at 05:30:34PM +0100, Lars Marius Garshol wrote:
> | One of the discussed syntaxes:
> |
> | <topic variable>[<type expression>;<conditions with simple paths>]
>
> Here I think the <type expression> should be merged with other types
> of conditions, since there seems little reason to consider it a
> different kind of thing.
Yes, a 'type constraint' is not so different from a 'condition
constraint'.
> In general I like the idea, though.
tolog is definitely missing a path sublanguage.
Yes, you can do everything with predicates, but, hey, yes, you
can also do everything with path expressions. Sometimes one is
easier and more natural, sometimes the other is.
I think one of the things tolog suffers from is that it is using the
'assembler approach' reducing _every_ kind of relationship to be
expressed as a predicate. This orthodox approach ...
> The main thing missing seems to be
> a consistent syntax and model for the path expressions that fits into
> the existing language. Something can probably be found, but I don't
> feel that it's happened yet.
... may make it hard to fit anything else in without compromising
consistency.
\rho
|
