Re: Blank Nodes Re: Toward easier RDF: a proposal"

> On 26 Nov 2018, at 07:26, David Booth < <>> wrote:
> On 11/24/18 6:50 AM, Hugh Glaser wrote:
> > And no, I don't want a Blank Node for [a location that has
> > labels] - the system that generates this should create a URI
> > if it doesn't already have one ;-)
> I like this line of thought.  I would much rather have auto-generated URIs, that are predictable and distinguishable as auto-generated, than blank nodes.  And even better, those auto-generated URIs could be generated using a standard algorithm, so that all tools would generate them the same way.
> As Aiden Hogen et all point out in "Everything You Always Wanted to Know About Blank Nodes": "the vast majority of blank nodes form tree structures", i.e., they do not contain blank node cycles.
> <>
> If blank node cycles were prohibited in RDF, then predictable URIs could be automatically generated for those blank nodes, bottom-up recursively based on the tree structure.  And prohibiting blank node cycles would not be a huge loss, because even the few cases that do use blank node cycles could be brought into conformance by replacing a few of the blank nodes with URIs, to break the cycles.

It also means that the problem of canonicalization/signature/etc would become way easier. The algorithms that I referred to in[1] are getting complicated due to those b-node cycles (I hope that Aiden, if he reads this, agrees with me). They would be way simpler if this restrictions was in place.


[1] <>

> David Booth

Ivan Herman, W3C 
Publishing@W3C Technical Lead
Home: <>
mobile: +31-641044153

Received on Monday, 26 November 2018 09:18:11 UTC