Re: how about a mix of GRDDL tutorial and use cases?

I agree with DanC - I think we should, as Ian pointed out and sort of
proposed at our first GRDDL meeting, separate out a tutorial from the
use-case document. The tutorial should be "for the people" (i.e. made
for someone that perhaps only understands HTML and an inkling of RDF) 
and be easy to read, along with links to working code and all the
details spelled out. Because of these constraints, the tutorial must be
simple and straightforward. I also think we should call the tutorial a
"primer" as to stick with W3C tradition.

    In contrast, the use-case document should offer exciting and diverse
use-cases of GRDDL. To this extent, I am much more comfortable with
having the wiki use case and XForms use case in the use cases document -
I just don't want them as tutorials. Yet I think they are important
because they reach out to exciting communities (Xforms, wikis, etc.) and
show how GRDDL technology is not just a "one-trick pony." So, while the
use-cases should be capable of being implemented with GRDDL (i.e. not
stretching GRDDL's capabilities too far beyond the original doc by Dom
and DanC [1]), I'm fine with these use cases being ambitious and not
implemented.

Then the reader who was new to GRDDL could read the tutorial/primer to
get a quick and simple application up and running, and then the reader
could read the use-case/scenario gallery and realize other cool ways
they can use GRDDL.

However, what use case should we use for the primer? I'm tempted to go
with something very simple, i.e. the booking a hotel/buying a
guitar/scheduling a meeting sort of use case.

                 cheers,
                      harry




Dan Connolly wrote:
> The use cases discussions are progressing nicely, but it
> occurred to me in post-meeting chat that our charter sets
> the expectation that an early deliverable will be a tutorial.
>
> There is a natural spectrum of cases, from
>
>  - things that are simple and straightforward, both
>    to learn and to implement, to
>
>  - more realistic use cases where we have implementation
>    experience, to
>
>  - use cases where we're pretty sure we know how we
>    would do it, to
>
>  - use cases that raise design issues that we haven't
>    figured out yet
>
> I suggest organizing our first publication along that
> spectrum. Let's start with a hands-on tutorial...
> something like hCard or dublin core and show how
> to add a link to a GRDDL transformation and use
> the results with SPARQL or OWL tools.
>
> Then in the next section, maybe something of the complexity
> of the hReview aggregation use case, where we'll have all the details
> worked out, but we might not spell out every step inside
> the document.
>
> Then on to stuff like educational metadata (and
> maybe clinical trial data; more on that separately)
> in wikis.
>
> And finally, cases like choosing one meeting out
> of a page of meetings, were we don't have the design
> completely worked out.
>
> What do you think, Fabien? Ian? Harry? Danny? Others?
>
>   


-- 
		-harry

Harry Halpin,  University of Edinburgh 
http://www.ibiblio.org/hhalpin 6B522426

Received on Thursday, 17 August 2006 17:29:57 UTC