[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [xri] Issue 1: Clarifying * Semantics
Actually, a GCS character is not a sub-segment per the BNF, but it certainly has the same characteristics as a sub-segment and is defined as a "qualified sub-segment" in the resolution section. If we view xri:@:1010 as having two sub-segments, there's no question that a) : is a sub-segment separator in xri:@:1010 and b) there's an implied reassignable sub-segment separator in xri:@1010. Dave > -----Original Message----- > From: Fen Labalme [mailto:fen@idcommons.org] > Sent: Thursday, July 08, 2004 11:11 AM > To: Dave McAlpin > Cc: xri@lists.oasis-open.org > Subject: Re: [xri] Issue 1: Clarifying * Semantics > > Colon is IMO *not* a subsegment separator in xri:@:1010 - rather, the > initial > GCS char is by definition a subsegment, and the colon is a decorator that > indicates persistence. Note that in xri:@1010, again there are two > subsegments, the GCS character and a reassignable subsegment. > > Fen > > > Dave McAlpin wrote: > > I was addressing Drummond's point about the difficulty of parsing. I > > just pointed out that the difference in complexity of an XRI parser is > > negligible between the current and the proposed syntax. Do you disagree > > with that? > > > > As to the clarification of : as a separator, I think there's a > > misunderstanding here. Colon is _always_ a subsegment separator. In > > xri:@:1010, there are two subsegments, @ and 1010. In xri:@example there > > are two subsegments, @ and example. Under both the current syntax and > > the new proposal, there's an implied * between @ and example in > > xri:@example. If that genuinely confuses users (which I don't think it > > does, it's simply an ignorable technicality), then we'd have to make it > > explicit as xri:@*example. That's what I was pointing out (somewhat > > facetiously) in my second comment below. > > > > Dave > > > > > >>-----Original Message----- > >>From: Fen Labalme [mailto:fen@idcommons.org] > >>Sent: Thursday, July 08, 2004 10:38 AM > >>To: Dave McAlpin > >>Cc: xri@lists.oasis-open.org > >>Subject: Re: [xri] Issue 1: Clarifying * Semantics > >> > >>Dave McAlpin wrote: > >> > >>>As we've pointed out before, the thing that makes parsing and > >>>interpreting XRIs difficult is cross-references. The difference > > > > between > > > >>>the two options we're currently discussing is negligible. > >> > >>While I agree that cross-references are the truly hard part, the > > > > proposed > > > >>simplification is *not* negligible. Rather, the change greatly > > > > clarifies > > > >>the > >>meanings of : and *. The first (an initial colon) becomes a simple > >>decorator > >>(I like that term) that denotes persistence - it is no longer > > > > *sometimes* > > > >>a > >>separator. The second (an initial star) now acts simply as a > > > > separator, > > > >>and > >>no longer must be implied in places that it does not appear. > >> > >> > >>>As for simplification of the rules, if implied * is confusing let's > > > > just > > > >>>require it. In other words, keep the current interpretation of * and > >>>change xri:@example/foo to xri:@*example/*foo, comparable to > > > > xri:@:3/:4. > > > >>>This is a much simpler change and has the benefits you mention > > > > below. > > > >>This would be dreadful. Don't forget that the reason we're changing > > > > from > > > >>dot > >>to star is for better human readability! > >> > >>Fen > > > > > >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]