Re: issues with RDB engines/dialects IRIs for EARL results

Hi Ted, Kingsley, others,

I'd like to continue this discussion.

On 19 Jun 2012, at 18:01, Ted Thibodeau Jr wrote:
> 1. Engines not listed above which should be added to cover all 
>   current WG member implementations/testing include Ingres, 
>   Progress OpenEdge, DB2, Informix, Sybase ASE, Sybase SQL 
>   Anywhere, OpenLink Virtuoso.  (We are testing against all 
>   of these.)  There may well be more.

I have added all of these to the table, and Firebird too as Kingsley mentioned it.

http://www.w3.org/2001/sw/wiki/RDB2RDF/SQL_Version_IRIs

> 2. There is a need to distinguish between versions of the above 
>   engines (e.g., Oracle 8i, 9i, 10g, 11i; Microsoft SQL Server
>   6, 7, 2000, 2005, 2008, 2012; MySQL 3, 4, 5; etc.), because
>   SQL:2008 is not supported by all current versions, never mind
>   all past versions, and R2RML/DM implementation support of 
>   these engines/versions will therefore vary.
> 
>   I don't *think* we need to go down to build/binary version
>   level, but I do think we need to go down as far as major
>   release version of each dialect/engine.

The obvious way of dealing with this would be to extend the product identifiers: rr:Oracle/8i, rr:Oracle/9i, etc.

>   This *could* be done by removing "Version" entirely from the
>   above ("Core SQL:2008" is arguably not a version but a dialect
>   of its own) and requiring a version disclosure along with the 
>   "Dialect" (which label I'd put in place of "Language Version" 
>   above)...

I have changed changed the table heading from "Language Version" to "Language Dialect".

(These IRIs are mostly intended for use with the R2RML property rr:sqlVersion to indicate compatibility of queries in R2RML view definitions. Hence the use of "version" here. I guess that rr:sqlDialect might have been more appropriate, but I guess that ship has sailed.)

> but that is potentially clumsy, and makes the simple
>   IRIs we've minted insufficient as identifiers.

I think the simple IRIs need to remain. I expect that many users will know that the SQL query they write for their R2RML view works on "whatever MySQL version I'm using", but they will not know or care about the specific version. So keeping the unversioned identifiers around is important. This doesn't mean we shouldn't do versioned ones too.

> 3. If we're going to keep a live results report page, then it
>   seems we'll need to keep a live identifiers page, such that
>   new engines can be added over time.

We certainly need to keep a live identifiers page. That's why it's on the wiki.

There are still two issues with this wiki page:

1. We need to move the page off the RDB2RDF wiki, because that wiki will be frozen when this WG ceases. We need to decide where it goes (W3C wiki?). We need to update the URL in the R2RML spec for REC.

2. I'm somewhat unhappy about having all these IRIs in the R2RML namespace. It might be better to put them into a different namespace. It's important to encourage the public to contribute to the wiki page, but this will inevitably lead to some level of clutter, and I strongly dislike the thought of encouraging people to clutter up a namespace that is defined by a W3C REC. (We placed rr:SQL2008 into the R2RML namespace, which is reasonable as it is the one dialect that the spec officially endorses. But that doesn't mean that other IRIs should go into the same namespace.)

3. Should we do something about JDBC and ODBC here? Kingsley said that they should be mentioned. I don't know enough about this. I'm aware of JDBC and ODBC as frameworks for connecting to all sorts of databases, but I'm not aware of JDBC and ODBC specifying dialects of the SQL language as such. Do they really do that? I thought the SQL dialect would still be the dialect of the underlying DB regardless of whether I connect to it through JDBC or ODBC or directly.

Best,
Richard




> 
> Thoughts?
> 
> Ted
> 
> 
> --
> A: Yes.                      http://www.guckes.net/faq/attribution.html
> | Q: Are you sure?
> | | A: Because it reverses the logical flow of conversation.
> | | | Q: Why is top posting frowned upon?
> 
> Ted Thibodeau, Jr.           //               voice +1-781-273-0900 x32
> Senior Support & Evangelism  //        mailto:tthibodeau@openlinksw.com
>                             //              http://twitter.com/TallTed
> OpenLink Software, Inc.      //              http://www.openlinksw.com/
>         10 Burlington Mall Road, Suite 265, Burlington MA 01803
>     Weblog   -- http://www.openlinksw.com/blogs/
>     LinkedIn -- http://www.linkedin.com/company/openlink-software/
>     Twitter  -- http://twitter.com/OpenLink
>     Google+  -- http://plus.google.com/100570109519069333827/
>     Facebook -- http://www.facebook.com/OpenLinkSoftware
> Universal Data Access, Integration, and Management Technology Providers
> 
> 
> 
> 
> 
> 
>

Received on Wednesday, 25 July 2012 20:09:03 UTC