Far McKon wrote:
> I think that would be cool, but man. That is super ambitious.
Once the schema is designed properly (I think I'm close) it shouldn't be
too difficult. Famous last words, I know. IMHO, the most difficult part
of this type of project will be converting a site like daviswiki over to
it. Site's like the one I'm most interested in (sacwiki) will be easy
because there's not much content yet.
> A good midpoint (to start with) might be something like:
>
> [DisplayMetadata()]
I've already got a macro that does this. This is a macro that gets all
metadata in the DB for the current page and displays it as an unordered
list.
> [MetaData("Key:=Value)]
Yes, the is the "set" function I was talking about. I hope to have this
working soon.
> I think the [[get(phone)]] and [[get(adddress)]] detaches the data yet
> another level/step from just typing crap in on some page,
You can still just type random data into a page. The metadata won't be
defined for that page though. I'm guessing those who care (ie the heavy
wiki editors) will spend a lot of their time cleaning up these types of
pages. But, imagine the time saved. Last time I checked the restaurants
page has restaurants listed by food type along with a phone number to
each restaurant. Imagine if one phone number changed... now imagine if
all of them changed.
> breaks the idea of simple markup,
Do you have any ideas on how to make the markup cleaner? I've simplified
it as much as I can see possible. Basically you have two bits of data
the name and the value. Everything else is done for you (including type).
> and can makes revision stuff harder.
> IE, now
> the contents aren't text, they are dependant on some other data some
> other place.
But that's just it. The contents are still just text. When comparing two
versions it might be cleaner to prevent the "get" expansions. Instead
just show the differences between where the metadata is set.
> the average non-cs joe might not 'get it' intuitively
This is where interface enhancements should come into play. If it's not
intuitive then we've done something wrong. Giving structure to data
isn't easy (how long has the term "semantic web" been around?) to do but
the power it gives us cs-types is vast.
> Overall I think this is nifty & useful stuff. Either way it is/should
> be easy to add some convince functions on top to make it easier or
> more intuitive if we need.
Yes, indeed. I imagine this feature will spawn many other features.
Scott
_______________________________________________
Sycamore-Dev mailing list
[EMAIL PROTECTED]
http://www.projectsycamore.org/
https://tools.cernio.com/mailman/listinfo/sycamore-dev
