DSML Teleconference Minutes July 25, 2001

 

 

1) List of attendees

 

Winston Bumpus, Chair  

Gavenraj Sodhi

Jeff Bohren

Tony Gullotta

Bill Curtin

Alexis Bor    

Jeff Parham

Andy Harjanto    

John McGarvey  

Keith Attenborough

Nigel Jacobs

Christine Tomlinson 

Rob Weltman, Palo Alto, CA    

Shon Vella

 

2) Minutes

 

 

Looking for editors

Website update

Individual presentations

Minutes from Chris Harding

 

 

Things to be added to Microsoft proposal

Things that need to be removed

Things to be changed

 

Editing of document

-------------------

 

Jeff Parham of Microsoft, Christine Tomlinson of iPlanet, and Tony Gullotta of Access360 will assist with co-editing.

 

Gavenraj Sodhi will take minutes.

 

Resolutions agreed to (e.g., 20010725)

 

Sequentially date these.

 

Action items (Track these)

 

--------------------------

 

Action:  E-mail presentations to group for posting to the website.

Winston Reviewed Summary of last week's meeting

 

Access360:  Unsolicited notification of change

      May set this aside because of time

 

iPlanet:  Goes hand in hand for DSML in asynchronous concept.  Still leaves the issue of unsolicited open.  Asynchronous is key element.

 

Microsoft:  Should throw out idea of Transport independence.  Need to get something out in 3 month timeframe

 

iPlanet:  If we take asynchronous approach, RFC2251.3 is transport independent, UDDI v1 and v2 are transport independent.

 

Microsoft:  Proposal on hand only discusses LDAP operations in DSML.  Should be able to overlay asynchronous on top of it.

 

iPlanet:  If we leave LDAP with collection of responses and operations, then we don't sacrifice transport independence.

 

Microsoft:  Would like more detail on this. 

 

Novell:  If a proposal can be drafted to show how this can be done, that would be best to show.  Buckets:  Add, change, delete, v3 bucket.  Would it be fine to write-up, iPlanet?

 

iPlanet:  Access360 proposal was in line of what we are talking about.  Would like to work on base proposal. 

 

Action:  Microsoft to send out base proposal in HTML and word formats.

 

Microsoft:  What is timeline?

 

iPlanet:  to be provided next week

 

Action:  iPlanet to provide updated modified proposal for consideration by next week.

 

Access360:  Filters are being passed as textural string.  We found it useful as nested XML tags.  The recipient found it useful to parse.

 

Novell:  proposal has comma separated list.

 

IBM:  not ready on interoperability comments:

 

Action:  will provide comments next week.

 

Agreement that security is out of band. 

 

IBM:  Parallel serial issue:  If we have batches, do we permit parallel as the advisory?

Should this be in DSML or LDAP?  There is a proposal in LDAP that may be implemented on the server.

 

iPlanet:  This might be substance for DSML v3.  Should be Tier 2.

 

Microsoft:  Seems important to supersede LDIF.  How can this bee done when a file has many operations in it.  Parallel vs. serial is that you don't want big entries in LDIF and perform them serially.

 

iPlanet:  You don't want to have 100,000 entries in XML to parse.  We should represent individual operations independently at a separate level where we can orchestrate separate operations.

 

Microsoft:  Interested in details.

 

Novell:  someone to write proposals around URIs and URLs?

 

VeriSign:  wasn't sure to write proposal on this.  Assumption was to wait before we decide on any one proposal.  Is definitely a transport issue.

 

Novell:  RFC 2255 LDAP RFC.

 

Microsoft:  Not required for referrals.  Still an open issue.

 

VeriSign:  transport issue unless you talk about referrals.

 

Microsoft:  referrals that don't depend on URLs.  URL naming does not come out in v2.

 

iPlanet:  URL would be present if another method was not defined.

 

Novell:  named URI

 

VeriSign and IBM:  Agree

 

Novell:  Referrals have a lot of further issues above and beyond.  Need to get hand on what it all means.

 

Placeholder placed on Referrals.

 

Novell:  Schema discovery and how people returned schema.  Should there be a standard format on schema representation in the DSML proposal?

 

Microsoft:  not sure if that was much of a question.  Do we need to transform aggregate schema that would be more useful to parse the large mess that comes out.

 

Novell:  Markup defines schema in v1.

 

IBM:  LDAP is trying to do some of this.  Suggestion is out to incorporate in schema.

 

Novell:  Can be expressed in DTD or schema (currently).  Sun agrees XML Schema.  Novell states XML schema as normative and than capture in DTD

 

Resolution (20010725):  To capture document as normative XML Schema and then capture in non-normative DTD.