Re: [xri] Proposal for XRI Syntax 2.1 to treat all delimiters as signficant

[Victor Grey <victor@idcommons.org>]

Drummond Reed wrote:
> Make XRI 2.1 syntax as simple and regular as possible, because we  
> cannot anticipate all future use cases for XRI syntax, and a  
> simple, regular syntax will better serve developers and  
> infrastructures that want to put XRI to uses we cannot yet anticipate.

Since we first started to show XRIs to groups of developers 3 years  
ago or more, there has been an ongoing struggle to simplify the  
syntax visually. The feedback we got from those early adopters was  
that they weren't going to be able to "sell" identifiers that were an  
over-the-top mess of weird looking punctuation - it undermines  
usability too much.

To engineers the visual "look" of these XRIs may seem like a trivial  
matter, and certainly not as important as unambiguous interpretation.  
But if we want to achieve widespread adoption, I can assure you all  
that it is an extremely important issue. We have done well with the  
basic "global i-name" syntax, but now we are faced with getting  
developer buy-in to a much more complex syntax necessary for data  
sharing.

I'm strongly supportive of the new proposals to make concatenated  
XRIs as clean looking as possible within the constraints of  
unambiguity, and I think it's well worth the extra effort on the part  
of the TC to hammer this out.

=vg