Re: Pretty printing OWL FSS : request for preferences and samples

Chris Mungall <cjmungall@lbl.gov> writes:
>>> If only annotations were the final argument rather than the first.
>>>
>>> Perhaps what we actually need is a new syntax designed from the ground
>>> up with specific requirements in mind
>>
>> I really hesistate with the idea that we need yet another syntax for
>> OWL. We do already have really quite a lot of syntaxes.
>
> Maybe some could be deprecated? It's not really clear who the target audience
> is in all cases and whether they're satisfied.

I think that would be a fine thing, although I am not sure how you could
find out who is using what.


> Of course the RDF flavors are inevitable and often useful, I think we all
> agree here that something that targets the OWL2 model directly is often
> required.

Indeed, I do agree.


>>> (e.g. layered on a sensible syntax like yaml, no lingering traces of
>>> lisp, VCSable, readable, moderately hackable in a text editor,
>>> incorporating @context ideas from JSON-LD).
>>
>> Of course, having said all of that, this was a large part of my
>> intention with Tawny-OWL. It fulfils all of the requirements above
>> except for the @context. And arguing the "lingering traces" -- it's far
>> from lingering in for Tawny.
>
> :-)
>
> Yes, but Tawny scripts aren't intended to be directly parsed by anything other
> than a lisp/closure interpreter.

Just so. So, a syntax and an associated evaluative semantics (defined by
tawny), as well as the OWL semantics. I'm not really holding Tawny up as
an alternative syntax (although, of course it is, and rendering for it
could be added to Protege). But, I used to think we needed a syntax that
did all of the things that you are talking about. Then I wrote tawny and
discovered that really, it wasn't a new syntax I needed after all.


> Whereas FSS is really just tweaked S-expressions isn't it?


If I remember correctly, FSS was just the closest thing to the abstract
syntax given in the OWL 1 spec. Something was needed that was easier to
type and read than XML. Not quite S-expressions, I guess, because they
put the symbol in the wrong place, outside the brackets!


> Though it's still practically the best of the bunch.
                                 
I rather like Manchester for reading, although I stopped writing it
because I found problems with the parsing.

Received on Thursday, 12 March 2015 17:00:12 UTC