[occi-wg] Fwd: Structured data over TCP?

Sam Johnston samj at samj.net
Wed May 27 14:49:02 CDT 2009


Tim's presentation is insightful.

Highlights (actually it's all good so this is most of the presentation):

   - Use plain text if... you possibly can.
   - JSON issues: Watch out for extensibility.
   - Use JSON if... The expected lifetime of the data is short.
   - Use XML if... your data is document flavoured or you're worried about
   i18n, extensibility or reusability.
   - Inventing New XML Languages:
      - Time-consuming.
      - Bureaucratic.
      - Difficult.
      - Unpleasant.
      - Includes complex software development as a sub-task.
      - Usually fails.
      - ... so try not to!
      - Some Good XML Languages
      - XHTML
      - DocBook
      - ODF
      - Atom
      - XMPP
      - UBL
      - RDF
   - Design Issue: Model vs. Syntax
      - “What matters is getting the data model right. The syntax is
      ephemeral.” vs
      - “The bits on the wire are the only reality.”
      - Design Issue: Minimalism vs. Completeness
      - “Let’s solve the whole problem.”
      - “Minimum progress required to declare victory.”
   - XML Schema Definitions (XSD):
      - Hard to understand, hard to implement, hard to interoperate.
      - One of the reasons why the SOA/WS-* project is sinking.
      - XML Extensibility: Three Options
      - No changes.
      - Must-Understand policy (e.g. as in SOAP).
      - Must-Ignore policy (e.g. as in Atom).

Morals of the story (IMO):

   - Bend over backwards to use plain text (if we can)
   - JSON is more interesting for RPC than ROA (Resource Oriented
   Architecture)
   - Avoid inventing new [XML] languages at all costs
   - 6 of the 7 "good" XML languages have already come up in the context of
   OCCI (NFI what UBL is)
   - Model and bits-on-the-wire are both important to us
   - RELAX NG and XPath is a GoodThing™
   - For XML extensibility ignore that which you don't understand (ala Atom)
   - Solve as much of the problem as we need to now but provide for
   extensibility later

Sam

---------- Forwarded message ----------
From: Tim Bray <tbray at textuality.com>
Date: 2009/5/27
Subject: Re: Structured data over TCP?
To: Patrik Fältström <patrik at frobbit.se>
Cc: apps-discuss at ietf.org


2009/5/26 Patrik Fältström <patrik at frobbit.se>:

> Do Apps Area have any general feeling about this?

I was going to start typing in a lengthy answer but then I remembered
I gave a tutorial at IETF70 on the choice of data formats for use in
Internet Protocols.  It's centered around XML but I think covers many
of the other issues that have come up in this.  Available for your
reading pleasure at http://www.tbray.org/tmp/IETF70.pdf

 -Tim
_______________________________________________
Apps-Discuss mailing list
Apps-Discuss at ietf.org
https://www.ietf.org/mailman/listinfo/apps-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/occi-wg/attachments/20090527/dd66788f/attachment.html 


More information about the occi-wg mailing list