- From: Dan Connolly <connolly@w3.org>
- Date: 06 Mar 2002 11:32:31 -0600
- To: Tim Berners-Lee <timbl@w3.org>
- Cc: Dan Connolly <connolly@w3.org>, www-archive@w3.org
Summary/TODO: 1) make a scehma 2) update http://www.w3.org/DesignIssues/RDB-RDF.html 3 ) link to and from coed 4) change bind to suggestPrefix(prefix, ns)? Hmm... "alpha-testing" or "peer review"? Anyway... Mar 06 10:40:22 <DanC> timbl? Mar 06 10:40:42 <DanC> export PYTHONPATH=/tux/u/connolly/lib/python:$SWAP Mar 06 10:41:28 <-- libby (~libby@yaddle.ilrt.bris.ac.uk) has left #swad (libby) Mar 06 10:45:05 --- DanC has changed the topic to: alpha-testing dbview... Mar 06 10:51:23 <DanC> use the same name for the users class and the document that gives you the columns in the users table? Mar 06 10:51:32 <DanC> TimBL: no... ones' a document and one's not. Mar 06 10:52:22 <TimBL> i think /administration/users is a document, a table schema Mar 06 10:52:48 <TimBL> i think /administration/users#zipcode is a concpet, a Property Mar 06 10:53:20 <TimBL> i think /administration/users#__Class is a Class, teh class of exactly things which are in the table Mar 06 10:53:26 <TimBL> i think /administration/users#__Class is a Class, teh class of exactly those theings which are in the table Mar 06 10:53:29 <DanC> see the docstring under class DBViewServer in http://www.w3.org/2000/10/swap/dbork/dbview.py Mar 06 10:54:51 <DanC> does this document vs. abstract-thing stuff apply to MIME types too? Mar 06 10:55:03 <DanC> TimBL: hmm... yes. er.. maybe not... Mar 06 10:55:31 <DanC> TimBL: we could say the mime type name denotes a registration document... Mar 06 10:55:43 <DanC> DanC: why can't we do that for database tables? the class is a document. Mar 06 10:56:33 <DanC> TimBL: I think I want to be able to refer to SQL table-classes as distinct from table-schemas. Mar 06 10:57:58 <DanC> see http://www.w3.org/2000/10/swap/dbork/dbview.py Mar 06 10:59:15 <TimBL> http://www.w3.org/2000/10/swap/dbork/dbview.py Mar 06 11:00:46 <TimBL> CLASS Mar 06 11:00:55 <DanC> bzzt. Mar 06 11:01:13 <DanC> __Class is much better Mar 06 11:01:24 <DanC> yes, we want an XML name that isn't an SQL name. Mar 06 11:01:53 <DanC> (and a property to relate the class with the class's schema document) Mar 06 11:02:13 <TimBL> suppose /admin db:table /admin/users#class Mar 06 11:02:59 <TimBL> ===> /admin/users#class db:hasSchema [ is log:racine of /admin/users#class ] Mar 06 11:03:22 <TimBL> ===> /admin/users#class db:hasSchema which has log:uri which is log:racine of /admin/users#class . Mar 06 11:03:48 <DanC> looks good; how would you like to write the http://www.w3.org/2000/10/swap/dbork/sqldb schema? Mar 06 11:04:21 <DanC> ... so I could take the @@ out of SwDB = Namespace("http://www.w3.org/2000/10/swap/dbork/sqldb@@#") Mar 06 11:05:16 <DanC> sink.bind('db', (RDFSink.SYMBOL, str(SwDB))) Mar 06 11:05:24 <DanC> should just be sink.bind('db', str(SwDB)) Mar 06 11:05:53 <DanC> TimBL: and bind() is a funny name... suggestNamespace() maybe? Mar 06 11:06:17 <TimBL> suggestPrefix() Mar 06 11:06:21 <DanC> yup. much better. Mar 06 11:06:34 <TimBL> suggestPrefix(prefix, ns)? Mar 06 11:06:39 <DanC> yup. Mar 06 11:06:53 <TimBL> *ACTION :x implement .-2 Mar 06 11:07:00 <DanC> btw... may I send a log of this chat to www-archive? Mar 06 11:07:12 <TimBL> yes Mar 06 11:07:15 <DanC> thx. Mar 06 11:08:06 <DanC> TimBL: did I already write the sqldb schema? Mar 06 11:08:10 <DanC> DanC: yup.. http://www.w3.org/DesignIssues/RDB-RDF.html Mar 06 11:08:29 <DanC> esp " Mapping an RDB into the Web - strawman" Mar 06 11:12:04 <TimBL> ... http://example:9003/dbview/inventory/products/1432 Mar 06 11:12:32 <TimBL> ... http://example:9003/dbview/inventory/products/1432 is a row defining document Mar 06 11:12:44 <TimBL> ... http://example:9003/dbview/inventory/products/1432/cost is a cell defining document Mar 06 11:14:43 <TimBL> ... http://example:9003/dbview/inventory/products/1432#cost may be the actual cost Mar 06 11:15:19 <DanC> for a cell, you had emps/rowid=123;col=shoe Mar 06 11:15:34 <DanC> is that a document? is this like the distinction between a document and a formula that is its contents? Mar 06 11:17:00 <DanC> RESOLVED: implementations should link to designissues notes and vice versa. Mar 06 11:18:18 <TimBL> WESHOULD 2) update Catalog Mar 06 11:18:24 <TimBL> scuse me Mar 06 11:18:33 <TimBL> 2) update http://www.w3.org/DesignIssues/RDB-RDF.html Mar 06 11:18:38 <TimBL> 3 ) link to and from coed Mar 06 11:18:41 <TimBL> 1) make a scehma Mar 06 11:19:01 <TimBL> swap/db Mar 06 11:19:18 <-- dom has quit (need network for my laptop setup) Mar 06 11:19:28 <DanC> 404 at http://www.w3.org/2000/10/swap/utils ; forbidden at http://www.w3.org/2000/10/swap/util Mar 06 11:20:55 <DanC> after all these details, the issue of UPDATE/POST looms large. Mar 06 11:21:48 <DanC> good... amaya didn't totally screw up newlines... lines: +11 -2 Mar 06 11:26:30 <TimBL> 4) change bind to suggestPrefix(prefix, ns)? Mar 06 11:26:42 <DanC> DanC: wanna alpha-test the HTML forms interface? i.e. getting data, not just the schema? Mar 06 11:26:45 <DanC> TimBL: not just now. Mar 06 11:28:34 <DanC> TimBL: when you spit out the schema, include count(*) while you're at it. Mar 06 11:28:40 <DanC> a table schema, that is. Mar 06 11:28:47 <TimBL> Size -- curent size of the table is a very useful ything to know up front for query opt Mar 06 11:29:16 <DanC> ADJOURN. -- Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Wednesday, 6 March 2002 12:32:30 UTC