realworldobject
-
@ PURL is operating correctly. Deref the Location URI to determine the Content-type of the target.
7:29 AM Jul 1st
from web
in reply to quoll
-
I'm looking for words suitable for the acronym GOOP. "GOOP Is Not A Protocol" (GINAP) is as close as I could get: . Help!
8:12 AM Jun 12th
from TweetDeck
-
@ If you want to stretch your LInked Data project from Atom to AtomPub, consider this blog entry:
9:49 AM Jun 8th
from TweetDeck
in reply to sgillies
-
The way to stream Linked Data using Web standards is to create an AtomFeed of Real World Objects. Clients can track down the RDF from there.
9:25 AM Jun 8th
from TweetDeck
-
@ My impl exp is that impls quickly become legacy millstones if you can't explain how its conceptual basis applies to new use cases
8:24 PM Jun 7th
from TweetDeck
in reply to sallamar
-
promotes unexpected reuse via a level of indirection between an RWO ID & extensible/context-sensitive/conneg info about RWO
8:05 PM Jun 7th
from TweetDeck
-
@ "Same" is a funny word. Some use cases would be undermined by "Samuel Clemens" owl:sameAs "Mark Twain". Others, not so much.
7:44 PM Jun 7th
from TweetDeck
in reply to edsu
-
@ Can you explain what you mean? Linked Data says that RWOs deserve unique names (URIs) that re/direct to URIs with info about the RWO.
6:01 PM Jun 7th
from TweetDeck
in reply to edsu
-
@ HTTP 301 reconciles different names for the same thing. HTTP 303 differentiates the same name for different things.
5:52 PM Jun 7th
from TweetDeck
in reply to jschneider
-
@ Let me put it another way. Every time we explain something to someone else, we almost inevitably name the concepts involved.
5:12 PM Jun 7th
from TweetDeck
in reply to jschneider
-
@ Chances are, the concepts are already named. Domain modeling reveals this: .
5:02 PM Jun 7th
from TweetDeck
in reply to jschneider
-
Real World Objects are the coherent, loosely-coupled concepts that deserve to be named. Domain models reveal RWOs relevant to use cases.
4:06 PM Jun 7th
from TweetDeck
-
REST: Everything is a resource, all operations are CRUD. Some resources are real world objects, some are info resources that describe RWOs.
3:45 PM Jun 7th
from TweetDeck
-
@ Linked Data is not rebranding . It is a recognition that RWOs are not info resources and deserve IDs. CRUD:
3:38 PM Jun 7th
from TweetDeck
in reply to dret
-
@ This obfuscated model helps explain why didn't catch on before now. httpRange-14 & clarify the model.
1:29 PM Jun 7th
from TweetDeck
in reply to dret
-
URI opaqueness: Encoding resource-type in a URI is a clue that the resource is well-modeled. Clients, OTOH, should NEVER take this on faith.
9:32 AM Jun 5th
from TweetDeck
-
An HTTP URI identifies a first-class object on the Web. A "service endpoint" is an incoherent, tightly-coupled first class object
8:31 AM Jun 4th
from TweetDeck
-
John Smith was onto something when he invented type-based surnames. Add DNS and you have semantically-potent globally-unique IDs.
8:15 AM Jun 4th
from TweetDeck
-
Enterprise architecture is to Web architecture as mainframes were to PCs in the 1980s.
8:02 AM Jun 4th
from TweetDeck
-
@ I wholeheartedly agree w/loose-coupling and REST. I guess I'm saying that evolution requires an intelligent designer. ;-)
6:06 PM May 15th
from TweetDeck
in reply to pkeane
|
- Name Jeff Young
- Location Dublin, OH
- Web http://q6.oclc.org/
- Bio Working on Linked Data: merging domain modeling, identities, and Web standards
|
