"Tag Triples" is basically a formalization of ad hoc triple systems.
Instead of the rigidity and thoroughness of ESW:RdfXml, you just say:
lion memberof communitywiki bayle memberof communitywiki mattis memberof communitywiki communitywiki isa group lion isa person bayle isa person mattis isa person communitywiki meetsat http://communitywiki.org/
Tag Triples has a website:
The basic idea is that when you interpret the file, you make it rigorous. This requires that you know what the file's being used for, where you found it, what the author means, etc., etc.,.
But it's something that's easy to play with, easy to think about.
There is a website for interacting with TagTriples:
You can do queries like: "Show me every title in the database."
select ?title where (?something title ?title)
Which results in:
title ongoing New Mac Boston ODF Day Meeting Hiawatha Shopping\, Dell vs. Apple RDF.net Is Expiring ...
Then if you wonder, "what are those things?" You can write:
select ?title, ?tag
where (?something title ?title)
(?something tag ?tag)And you get something like:
title,tag ongoing,feed ongoing,_:xmlnode-5-1 New Mac,entry New Mac,_:xmlnode-5-14 Boston ODF Day,entry Boston ODF Day,_:xmlnode-5-82 Meeting Hiawatha,entry ...
Here we discover that there's something called an "entry."
Let's look for just entries:
select ?title
where (?something title ?title)
(?something tag entry)And we see:
title New Mac Boston ODF Day Meeting Hiawatha Shopping\, Dell vs. Apple RDF.net Is Expiring A Pepper By Any Other Name WS-Stardate 2005.10 ...
Okay, this is cute- but how do I learn what other things there are that ?somethings could tell us?
To do this, we need a graph browser. The only way I know to get to the browser is by the search function:
Search for one of the titles we found, such as "Boston ODF Day", and then look at the data the browser gives you.
Here's a URL that jumps right there:
There, we can see that there's an "id" field, which is the URL of the thing.
select ?title, ?id
where (?something title ?title)
(?something tag entry)
(?something id ?id)And we get:
title,id New Mac,http://www.tbray.org/ongoing/When/200x/2005/10/29/New-Mac Boston ODF Day,http://www.tbray.org/ongoing/When/200x/2005/10/28/Boston Meeting Hiawatha,http://www.tbray.org/ongoing/When/200x/2005/10/28/Hiawatha-Bray Shopping\, Dell vs. Apple,http://www.tbray.org/ongoing/When/200x/2005/10/26/Dell-Huh RDF.net Is Expiring,http://www.tbray.org/ongoing/When/200x/2005/10/24/RDF.net A Pepper By Any Other Name,http://www.tbray.org/ongoing/When/200x/2005/10/24/Rose-Pepper-Capsicum ...
So, it's pretty cool!
Now, we can load in the data that we had above.
lion memberof communitywiki bayle memberof communitywiki mattis memberof communitywiki communitywiki isa group lion isa person bayle isa person mattis isa person communitywiki meetsat http://communitywiki.org/
You go to the post-data page, …
And paste that block in. (The "lion memberof communitywiki" yadda yadda yadda one.)
Tell it that the format is tagtriples, in the little combo box on the page.
After you've loaded it in, you can confirm it's there, on the graphs page.
Perhaps it's not there any more, but when I did it, it had placed the data at:
(Again, it may be gone by now.)
Now we can put in a query:
select ?x where (?x memberof communitywiki)
…and we get back:
x lion bayle mattis
W00t!
I think this would be easy to scrape from a wiki.