RE: [xri] GCS Characters

John’s right - =drummond@microsoft.com is just as valid an XRI as @microsoft.com=drummond, and has the same type of global inference. It’s a reference to the organization in the context of the individual, and I can think of lots of uses for such a reference.

However I think Steve’s original concern is no longer an issue under the new XRI-as-relative-URI architecture, because the XRI bound to a base URI is no longer going to be confused with an email address. Examples:

http://xri.net/@microsoft.com=drummond

http://xri.net/=drummond@microsoft.com

info:xri/=drummond@microsoft.com

In fact, even an ordinary email address when included in the path of a (non-mailto) URI is no longer an email address – it’s just a path.

http://xri.net/drummond@microsoft.com

=Drummond

From: John Bradley [mailto:jbradley@mac.com]
Sent: Sunday, November 23, 2008 4:15 PM
To: Markus Sabadello
Cc: OASIS XRI TC; David Orchard
Subject: Re: [xri] GCS Characters

You might want to do it for some malicious reason.

If we don't preclude it, the syntax is valid stupid or not.

If @microsoft.com=drummond is a subject in Microsoft's authority server that it is inferring has something to do with =Drummond then why wouldn't =drummond@microsoft.com be a subject in =Drummond's authority server hat he is inferring has something to do with the global @Microsoft.com

If we allow it someone will use it and if it can be misused someone will.

Steve's quite legitimate question is will the benefits of saying =drummond@Microsoft.com vs =drummond*(@microsoft.com) outweigh the possible social engineering dangers that are introduced.

We need a clear answer to that. The answer may well be yes they do, but we need to consider it carefully.

John B.

On 23-Nov-08, at 3:58 PM, Markus Sabadello wrote:

Why would you ever want to put a company into the context of an
individual (what does =drummond@microsoft.com mean??). To me it makes
sense only the other way round, i.e. @microsoft.com=drummond.

And something that never appears anywhere can hardly confuse people, or? :)

Markus

On Sun, Nov 23, 2008 at 7:25 PM, John Bradley <jbradley@mac.com> wrote:

I spent some time on the phone yesterday with Steve Churchill reviewing the

proposal on GCS symbols in XRI 3.0.

He reminded me that one of the original objections was allowing:

=drummond@microsoft.com

http://xri.net/=drummond@microsoft.com

This in 2.0 would be represented as:

=drummond*(@microsoft.com)

http://xri.net/=drummond*(@microsoft.com)

In removing the need for parenthesizing cross-references we run the risk of

people misconstruing direct concatenated XRI as email addresses or otherwise

trying to read the first XRI segment from right to left.

The problem is alleviated slightly by having the XRI bound to a scheme.

This is a social engineering problem rather than a actual resolution issue.

Thoughts?

John B.

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

To unsubscribe from this mail list, you must leave the OASIS TC that

generates this mail. Follow this link to all your TCs in OASIS at:

https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php

xri message