Re: [limit per resource] implementation experience and/or concerns?

On Friday 17 April 2009 14:15:24 Steve Harris wrote:
> Hm, interesting. The wiki page gives SELECT examples, and it's a  
> little hard (for me at least) to imagine what it would look like in  
> DESCRIBE land. Is it something you could do in CONSTRUCT? 

Ah, good point. I added a CONSTRUCT to the wiki some time ago, but I've added 
a DESCRIBE example now too:

http://www.w3.org/2009/sparql/wiki/Feature:LimitPerResource#Use_Cases
(would this rather belong in the examples section?)

So, basically, it isn't very natural to think in terms of solutions in many 
cases. Importantly, you seldom have a clue how many solutions you expect, 
since some properties may or may not exist, there may be many triples that 
share subject and predicates, but not object, and you have no idea how many 
different objects there are. All this makes the number of solutions 
uninteresting, and often a LIMIT on number solutions is impossible to compute 
in advance and therefore totally useless. 


> Or do you  
> have a lot of per-schema smarts in your DESCRIBE implementation?

Nope, plain SPO here :-)

> Agreed, but there are parallels. Subqueries + aggregates in SQL can do  
> something like this.

I guess so...

Anyway, I have a +1 on subqueries, so by all means, I hope we can get this as 
a easy shorthand for subqueries, but I'd like to have it anyway.

Kind regards 

Kjetil Kjernsmo
-- 
Senior Knowledge Engineer
Mobile: +47 986 48 234
Email: kjetil.kjernsmo@computas.com   
Web: http://www.computas.com/

|  SHARE YOUR KNOWLEDGE  |

Computas AS  PO Box 482, N-1327 Lysaker | Phone:+47 6783 1000 | Fax:+47 6783 
1001

Received on Friday, 17 April 2009 14:51:19 UTC