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

# xdi message

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

Subject: Re: one question about

• From: Drummond Reed <drummond@connect.me>
• To: Markus Sabadello <markus.sabadello@xdi.org>
• Date: Wed, 12 Sep 2012 14:23:34 -0700

Thanks Markus for clarifying the question. It's late for me in London so I can't answer now, but let me cook on it and see if I can post something tomorrow.

=Drummond

On Wed, Sep 12, 2012 at 7:51 AM, Markus Sabadello wrote:
Without cross-references, this is easy to answer.

I.e. the following:
=markus/+friend/=animesh

and the following:
()/()/=markus
()/()/=animesh
=markus/+friend/=animesh

Are just different ways of serializing the same graph.

But I am not so sure when it comes to cross-references.

I.e. are this
=x/+y/(=markus/+friend/=animesh)

and this the same graph?
=x/+y/(()/()/=markus)
=x/+y/(()/()/=animesh)
=x/+y/(=markus/+friend/=animesh)

With my current understanding of the graph model, I'd have to say no, they are not the same.

Generally I think the "test" whether two graphs are equals is to draw them.

Markus

On Wed, Sep 12, 2012 at 12:19 PM, Drummond Reed wrote:
Markus, my sense is that this question is entirely about the question of equivalence of XDI graphs with and without explicit context statements.

My gut reaction is that the TC should clearly state the rules for XDI graph equivalence on way or the other, i.e., by declaring non-essential context statements either MUST be included or MUST NOT be included in graph comparison, and that should solve the issue.

Then the serialization formats should key off those same rules.

Which way are you leaning, i.e., do you think non-essential context statements MUST or MUST NOT be considered significant?

=Drummond

On Tue, Sep 11, 2012 at 1:19 PM, Markus Sabadello wrote:
Drummond and others,

If I put the following into the XDI Converter and try to convert to JSON:

@!9999!8888\$msg\$(!1234)\$do/\$add/(()/()/=markus)
@!9999!8888\$msg\$(!1234)\$do/\$add/(=markus/+friend/=animesh)

Should the output be this?

{"@!9999!8888\$msg\$(!1234)\$do/\$add": [{
"()/()": ["=markus"],
"=markus/+friend": ["=animesh"]
}]}

or should it be this?

{"@!9999!8888\$msg\$(!1234)\$do/\$add": [{
"=markus/+friend": ["=animesh"]
}]}

Or are both valid conversions? Or does it depend on the "contexts" parameter?

If both outputs are valid, that would imply that both represent the same graph, which however does not seem very intuitive in the XDI DISPLAY format.

Prashant, Pankaj and I ran into this while updating the XDI/JSON implementation.

This issue reminds me of a new idea involving () arcs and inner graphs that you mentioned to me, perhaps it's time to explore this further...?

Markus

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