OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

sca-assembly message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Issue 87: My issue with saying that definitions is a single infoset

Infoset [1] is an abstraction of the XML document/model so that there is
no dependency on the serialized aspect of XML.

Each definitions.xml file (and there can be many in the domain) have a
document root element:

<definitions xmlns="http://docs.oasis-open.org/ns/opencsa/sca/200712";


The target namespace applies to all the children (for example intents)
of the root element.

If we say that definitions is a single Infoset (within the domain), that
implies that there is [document element] property that contains a single
element information item with a [namespace name] property of
"http://docs.oasis-open.org/ns/opencsa/sca/200712"; and a [local name] of
  "definitions". This element information item has a single attribute
information item in its [attributes] property with an empty value for
[namespace name] and a [local name] of "targetNamespace".

Given this I do not see how definitions can be a single infoset in a
domain. It is really a collection of definitions elements and since
Infoset models XML, it is never a collection but a hierarchy. The only
way we can have a single infoset is if we create a synthetic infoset
from the various definitions files in the domain and have a synthetic
[document element] property with a [local name] of "domain-definitions"
(for example).

Either we have to say that it is a collection of definitions that are
globally applicable and there should be no conflict. OR we say that it
is a single Infoset and define how the synthetic single Infoset is
constructed from multiple definitions files with a single [document

In looking at our spec, I don't see any place where it says that
definitions is a single Infoset within the domain. The only place where
the spec talks about Infosets is in section 1:
"This specification is defined in terms of Infoset and not in terms of
XML 1.0, even though the specification uses XML 1.0 terminology. A
mapping from XML to infoset is trivial and it is suggested that this is
used for any non-XML serializations."
This, in conjunction with section 9, says (to me) that there are
multiple definitions Infosets within the domain not a single Infoset.

I know we discussed this (definitions being a single Infoset in the
domain) during the Boston F2F. So not sure if this is a result of a
particular issue resolution not being captured correctly or applied
correctly. But as far as section 9 goes, it does *not* say that there is
one definitions Infoset per domain, so I'm happy with the spec as is and
can't propose a resolution (other than support CNA)


PS: after re-reading my email, I think there may have been a disconnect
between Mike and I. I was talking about a single definitions Infoset
*within the domain* and how that does not work. This came about as a
result of issue 44 resolution which was to allow multiple definitions
files but have a single Infoset. I'm wondering if Mike was talking about
single definitions Infoset *per contribution*.

[1] http://www.w3.org/TR/xml-infoset/

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]