Re: New proposal: health & medical extensions to schema.org

On Tue, May 15, 2012 at 5:16 AM, Dan Brickley <danbri@danbri.org> wrote:

> On 15 May 2012 00:15, Thad Guidry <thadguidry@gmail.com> wrote:
> >
> > 1. Is this
> >
> > <span itemprop="name">Dr. John Smith</span>
> >
> > on the MedicalScholarlyArticle example
> >
> > going to cause Parsing problems ?  Use of the short form "name" of the
> Author conflicting with the short form "name" of the
> MedicalScholarlyArticle within the same ?
>
> In both http://schema.org/Person and http://schema.org/CreativeWork
> classes, we see 'name' listed as 'Properties from Thing'.
>
> In http://schema.org/Thing (the most general of all classes), the
> property 'name' is defined as 'The name of the item.'
>
> I believe that covers the name of author/person case, and the name of
> creative work (incl. scholarly article, medical scholarly article,
> etc.) too.
>
>
>
Yeah, I know that.  I was more worried about a parser going... Ah, OK,
here's the name of the the MedicalScholarlyArticle...oh wait, no, HERE's
the name of the MedicalScholarlyArticle...oh wait no...THIS is the name of
the MedicalScholarlyArticle...etc...

Our current documentation around how to "READ" through Microdata coded with
Schema.org is fairly lackluster.  This page
http://schema.org/docs/gs.html tells
a bit about itemscope and where 1d. Embedded items shows nice closing tag
syntax, but what should a "good" parser do when it meets up with potential
snafus were the itemscope might unintentionally continue because of missing
a </p> or a missing </div> etc.   That's where my worry was founded just
quickly scanning the markup in the MedicalScholarlyArticle example and
building a parser in my mind.

We tell how to markup with Schema.org pretty well, but not about reading it
back properly.  Perhaps someone has begun documentation on that level and I
do not know about it ?


> > 2. And where's the ending </p> in the example ... hidden further and
> just expected ?
> >
> > I basically worry that the combination of both of those in the example
> might cause some shortsightedness with any Schema.org parsers themselves.
>  Is my worry unfounded or will itemscope itself keep everything in check
> even if there's 1 <div> ?... because there are a whole lotta "name"
> properties in that 1 example and the scope itself keeps changing through
> the <div> and stacking along with <span> elements instead ... scary stuff
> to me.
>
> Looking at the example, it might be clearer if <p> was opened and
> closed, and if the closing of each <span> was on its own line, nested,
> to show the structure.
>
> But yeah, real-life markup will be chaotic and often broken. As far as
> I can see, the example says what it (hopefully) wants to say. It is
> easy to get things wrong though, and that's where better tooling and
> checkers will help us all.
>
> BTW I was recently talking with Manu Sporny and others from the RDFa
> group (RDFa being a close relation of Microdata syntax) about such
> tooling, and Manu took some rough prototypes and turned them into
> quite a nice tool. Currently it only does RDFa but we could imagine
> easily enough having a microdata parser included too:
> http://rdfa.info/play/  ... perhaps something like that (where you can
> edit examples directly, and *immediately* see the corresponding
> extracted data) could help us all get this stuff right? Whenever
> experts get confused, we have to think "what'll this be like for busy
> webmasters...?".
>
>
Oooo... yeah, that's a very, very cool tool... Manu get your "geek" on and
hack that thing for Microdata already ! :)


> Anyway, the example looks ok to me. I'll try converting them to RDFa
> and put them in this tool, to see how things look that way.
>
> cheers,
>
> Dan
>



-- 
-Thad
http://www.freebase.com/view/en/thad_guidry