[foaf-dev] beyond foaf:mbox_sha1sum
henry.story at bblfish.net
Sat Dec 19 22:37:38 CET 2009
On 19 Dec 2009, at 21:19, Gregory Williams wrote:
> n Dec 19, 2009, at 4:13 PM, Richard Cyganiak wrote:
>> Hi Greg,
>> On 19 Dec 2009, at 19:44, Gregory Williams wrote:
>>> I'm open to the idea that smushing can lead to unwanted results,
>> That's not just an idea, it's a well-documented phenomenon that multiple parties who have worked with FOAF crawls have encountered and documented.
> I'm just trying to make the point that this seems to be reason to get the people emitting bad sha1sum triples to fix their code, not to scare people off of using sha1sum when it's legitimate.
Smushing on the linked data web is just not as simple as it seems to OWL reasoning systems.
OWL reasoning is based on how to smush two graphs where one believes both to be true.
But on the linked data web, graphs can be partly broken, completely wrong, lying, etc,...
So what does one do when one comes across two graphs that have mistakes in them? One can reject them completely, or one could go and find the smallest change required to that graph to make it less incompatible with the rest of one's information.
That is more the type of reasoning that linked data browsers will need to do.
This does not mean that OWL is wrong. If indeed the sha1 sum of an mailbox is inverse functional one can write it. The problem is that one just cannot merge all graphs on the web.
More information about the foaf-dev