Re: Comments on RQ25

Orri,

Orri Erling wrote:
> SPARQL RQ 25 Notes
> 
> 
> Here are some initial comments.
> 
> While reading the document, I found the following typos:
> 
> 
> dervived

Can't find it - which version are you reading?

> equivilent

Fixed (Eric in sec 11 in case you have a separate copy)

> displacment

Fixed (Eric in sec 11 in case you have a separate copy - data was wrong but 
not the query in 11.4.10bis)

> 
> 
> At the start of 4.4, readability could be better.  The casual reader gets
> confused about basic and group graph patterns.   Since I had not myself
> written our  SPARQL parser, I had to go to the grammar to see the point.

Added a note to do something about this so the point isn't lost.

> 
> It should be said somewhere at the start that:
> 
> Basic graph pattern consists  of zero or more triple patterns.  For the
> pattern to match, all vvvariables and blank node  labels must be bound so
> that triples corresponding to the bindings exist in the data.  If a basic
> graph pattern is empty,  it has one solution and produces no bindings.

Added a note so this does not get lost.


> Also, the point about the scope of a blank node label should be reiterated a
> few times in the text, something like:"
> 
> {{_:b1 foaf:name "Alice"} {_:b1 foaf:mbox ?box}}
> 
> is not the same as
> 
> {_:b1 foaf:name "Alice" . _:b1 foaf:mbox ?box}
> 
> because the scope of a blank node label does not go outside of a basic graph
> pattern.

If we go with the proposal arising from the last telecon, we add:

4.1.4
"""
The same blank node label may not be used in two separate basic graph patterns.
"""

and also I've added a new "A.6 Blank Node Labels" so this can also be put near 
the grammar.

> Orri
> 

	Thanks for the review,
	Andy

Received on Friday, 26 January 2007 14:05:49 UTC