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

 


Help: OASIS Mailing Lists Help | MarkMail Help

xri message

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


Subject: RE: [xri] Issue 1: Clarifying * Semantics


Dave and I talked late today about this issue (he's
headed out of town tomorrow), and I think our
conversation helped me better understand the crux of
the issues here. In this message I'll try to: 

a) summarize that understanding first, 
b) clear up one misconception,
c) further clarify one point being discussed, and
d) share a long-term perspective that I have been
applying to this issue.

SUMMARIZING THE ARGUMENTS

The case for "dual subsegment delimiters" (star and
colon both) seems to rest primarily on two points:

1) Backwards compatability with XRI 1.0.

2) Compactness of consecutive persistent XRI
subsegments because only colon is needed to both
delimit and "decorate" them.

The case for "one subsegment delimiter and one
persistence decorator" also seems to rest primarily on
two points:

1) Simplified parsing and interpretation rules (for
both machines and humans)

2) Freeing up colon to be used by producer-specific
algorithms (see new subthread I just posted).

CLEARING UP ONE MISCONCEPTION

In messages earlier today there were references to my
early proposal to do away with the concept of
subsegment altogether by putting star on an equal
footing with slash. Although I still believe in the
underlying motivations for this proposal, I was
convinced by the subsequent discussion that it was
better to keep the current "two-level" structure of
XRIs, i.e., major segments (slash delimited) and minor
subsegments. The reason was backwards-compatability
with URIs, which depends heavily on having only a
single major segment delimiter (slash).

However the main motivation behind this proposal was
and still is simplification of XRI syntax. The simpler
the better (as long as it is still sufficiently
expressive). To that end...

FURTHER CLARIFYING THE SIMPLIFIED
PARSING/INTERPRETATION RULES

Since one of the main motivations for the single
subsegment delimiter approach is the simplified
parsing and interpretation rules for both machines and
humans, I thought it would be helpful to actually list
the rules involved. My inspiration for this was my
experience updating the proposed XDI addressing rules
to reflect single subsegment delimiter syntax. I found
myself very surprised at how much simpler things got.

So, herewith my summary of the rules involved:

RULES FOR DUAL SUBSEGMENT DELIMITERS:
Rule 1) Slashes delimit segments
Rule 2) Stars or colons delimit subsegments UNLESS the
star or colon immediately follows a slash, GCS
character, or opening parentheses, in which case the
star or colon is only a reassignable or persistent
decorator
Rule 3) A segment or subsegment is persistent if it is
preceeded by a colon used EITHER as a delimiter or a
decorator
Rule 4) A segment or subsegment that is preceeded by a
star that is NOT used as a delimiter is equivalent to
the same segment or subsegment without the star

RULES FOR A SINGLE SUBSEGMENT DELIMITER:
Rule 1) Slashes delimits segments
Rule 2) Stars delimits subsegments
Rule 3) Colon decorates (prefixes) any persistent
segment or subsegment.

I don't know about anyone else, but I find the
differences in these two rulesets striking. I think
the reason is that the whole concept of a
"reassignable decorator" just goes away. It's not
needed anymore. Delimiters are delimiters, decorators
are decorators, and only one decorator is needed -
colon.

TAKING THE LONG-TERM PERSPECTIVE

In several other recent conversations I have shared
the perspective I've been applying to this issue,
which is: "What is best thing for XRIs in the long
run?" Yes, the issue of backwards compatability with
1.0 is very important. But realistically, the XRIs
that exist today are probably less than .001% of the
XRIs that will exist 10 years from now. We are at the
very earliest stages of real-world deployment.

Thus, as we have agreed several times, 1.1 is our
single best opportunity to make any changes to the
syntax that make it as clean and simple as possible
for the 99.99% of all XRIs to come.

This is the reason I am a proponent of the single
subsegment delimiter proposal. It is in my view a
fairly dramatic improvement in simplicity. When paired
with the advantage that it frees up colon to be used
(along with dot) in producer-specific algorithms for
subsegment composition - and that this cushions the
problems with backwards compatability - I believe the
benefits well outweigh the costs.

=Drummond


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