RE: LANG: first sketch & TEST abstract syntax

> So the idea is that we specify the language using some sort of
> "abstract" (and hopefully more compact) syntax, then define mappings
> from that into various "serialisations" (RDF, XML etc.).

Personally I am supportive, and hope that the TEST group can, if necessary,
initially specify some tests in this abstract syntax, which later on will
get translated into the preferred concrete serialization.

Obviously in an abstract syntax we can't automate a test, but at this stage
I think that that is not the key goal of developing the test cases.

In contrast with Frank's concerns I do not think the test area we should put
a hard dependency the lang area coming up with a concrete syntax.

We currently seem to be using DAML+OIL, and I guess the assumption is that
this can act as an alternative concrete syntax for the abstract syntax that
is yet to be agreed ....


Received on Friday, 8 March 2002 09:58:17 UTC