[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: minutes from ER TC meeting 20030422
Present: Norm, Paul, Lauren, Ardo, Mark, Normand Issue: about interleaving TR9401 and XCatalogs. If an XCatalog points to a TR9401 catalog, the implementation is allowed to understand the format (as it is allowed to understand any format it wants), but it is not mandated, even if the implementation understands standalone TR 9401. We believe the spec reflects this. Lauren to get back to the commenter about this. Issue: is a recommended way to handle external identifiers when parsing catalog files themselves? Norm is not sure we can add something to help bootstrap. Perhaps we could add a non-normative note saying the problem exists and implementors should think about it. The spec is currently silent on this point. If you attempt to load the catalog and it has a doctype declaration, how do you fetch the DTD in the first place? Is there some sense to having an implementation try to fetch that? There does not appear to be any benefit. You may not need to try to resolve that first doctype through the catalog. Users should note that this may not be resolved. The first catalog loaded may have a system ID that has http but you may not be connected to the internet; in this case a catalog resolver should be allowed to figure out what to do itself to allow the process to continue. Potentially there is a little bit of magic here in the implementation. Are the URIs in the catalog subject to catalog resolution? Are the nextCatalog etc subject to catalog resolution? Is the RHS of the catalog entry subject to catalog resolution? In general the RHS is not subject to resolution. Otherwise you can get into infinite loops. Issue: copyright and licensing for the DTD and schemas: Lauren action item to find out the OASIS policy on this stuff and come up with recommendations. Issue: http://www.oasis-open.org/archives/entity-resolution- comment/200304/msg00012.html This will be made more explicit in the spec. The resolution is as stated by Norm in that email. No other issues with the current draft. Norm will add the three notes (magic/bootstrap, RHS are not subjected to resolution, system identifiers are not made absolute before being passed in to resolution). Norm will do that soon so we can go back to committee spec. Lauren to get the Debian implementation added to the implementation list on the group page, and the rest updated. Implementation guidelines: Mark has finished the first draft, based on a hierarchical system implemented at Debian. Question: people can give aliases to chunks of the catalog. The catalog resolver makes the alias point to the right place. This is not in the spec, but may be easier for users. Mark will post the first draft for us to comment on. Next meeting: April 29th, 11 am PDT, 2 pm EDT. Paul sends regrets. Lauren -- Lauren Wood, Textuality Chair, XML 2003 http://www.xmlconference.org
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]