Re: comments on section 7.4

Hi Carlos,


>
>     1.I am not sure we can avoid any technology-specific mentions in the parts out of implementation approach. This section is about difficult notions, and I believe examples throughout the section will help readers. Of course we should try to pick examples from representative technologies, not just LD.
>
>
> We should try and rethink about any of those where we will not be finally able to achieve it, because that won't be a good signal. Technologies change and evolve and new ones will also arise. If we do a good work BPs will remain (see WCAG 2 example and how they solve exactly that problem they had with WCAG 1)


The problem is that any effort to unify terminology for these things is likely to fail - or to result in the next new reference terminology, which will be as in
http://xkcd.com/927/ (just replace 'standard' by 'terminology')
I prefer the bottom-up approach, to rely on examples in existing technology. Not perfect.
Anyway I don't want the discussion to happen. Just warn that I'm not expecting a brilliant, all-issue-solving outcome.



>
>     2. I don't know a lot of recipes for implementing these BPs outside of the LD realm. And I actually doubt it will be easy to find. I mean, LD technology has been made to give technical options and best practices to solve these issues, hasn't it? If other technology had been particularly good at it, the need for LD would have been less good... But of course I eager to learn about any solution elsewhere!
>
>
> Well, as a first shot of examples:
>
> BP11: A good documentation could be just a PDF, a UML model or XML schema, etc. as well.


PDF is fine.
UML model is already hinted by the diagrams, but we can add more to the BP indeed.
XML Schema would require to specify which feature should be used to indicate definitions (otherwise it's just like an OWL file without rdfs:comment attached classes and properties)


> (note that at least as currently the BP does not require documentation to be machine-readable)


The text is not machine-readable, but the way it is attached to the vocabularies (rdfs:comment, skos:definition, etc) is. Which is different from a PDF.


>
> BP12: There are plenty of examples for this outside LD: (HL7-set, DICOM, GELLO, CCOW, UBL, OCDS, HR-XML, SDMX, GML; KML; SLD; WCS; WFS; WMS; XBRL...) I think we just need more diversity in the examples here.


Agreed.


>
> BP13: Regardless terminology I think this is already quite neutral
>
> BP14: Only more variety of examples may be necessary here e.g. https://sites.google.com/site/erwinfolmeronsemanticstandards/list-of-semantic-standards or http://www.ssi.dk/graphics/standardkatalog/2.0/index.html (and terminology)


Yes these sort of things would be helpful, though I expect we'll have to do more homework to clarify the status of these things we would throw at readers. and pick more official things than Google Docs!


>
> BP15: This is without doubt the most problematic as it highly biassed as currently (and also being discussed even from the LD perspective). So I think we may need to rework/transform/drop it.
>


Actually your comment makes me think: what if LD badly needs a best practice, which other technology would need less? Should we refrain from putting it?

Cheers,

Antoine

Received on Friday, 23 January 2015 16:31:26 UTC