RE: [xri-editors] Reviewing RC1b

["Dave McAlpin" <dave.mcalpin@epokinc.com>]

Comments on RC1b

There's a problem with the index - for some reason Word thinks there's a
section 3.1.1.1

The note in the definition of XRI reference about fragment identifiers seems
like a non sequitur.

Section 2.1.1 has a broken reference on line 629.

The example added to section 2.2.1 is good, but I think we should strike
"Further examples are provided" sentence because it refers to a particular
version of a draft, which we've tried to avoid at least with respect to IRI.
The only other place we refer to a specific section is in 2.2.5, and I think
we should probably drop the section reference there as well.

The last sentence of 2.2.4.1 implies that there's a potential security
problem with $!. It doesn't elaborate but instead refers to section 4.
Section 4 doesn't contain any info about this issue that I could see.

I'm not sure the new sentence in step 1 of 2.2.4.3 that says "This places
the XRI into escaped normal form" is completely accurate. Escaped normal
form, in my mind at least, has more to do with proper escaping than it does
with UTF-8 encoding. I think it might be more correct to say something like
"Escape characters as necessary to place the XRI in escaped normal form as
defined in section 2.2.4." Or just assume it's in escaped normal form to
start with and leave that sentence out altogether.

In 2.2.4.4, the reference to section 2.2.4.3 didn't print correctly for me
(appeared as a little box). It looks ok inside Word, but may be worth double
checking.

Section 2.4.1 says it's going to describe canonicalization and define the
rules for putting an XRI into canonical form, but instead it describes
equivalence rules. Rules for creating a canonical form should be things like
"drop the leading dot" and "remove cross-references that begin with a !".

Appendix B asks if $! is sufficiently specified by 2.2.4.2. I don't think it
is. In fact, this section seems a little terse overall. I'd hate to have to
implement to Appendix B if all I had to work from was the spec itself.

Dave

> -----Original Message-----
> From: Drummond Reed [mailto:drummond.reed@onename.com]
> Sent: Monday, November 10, 2003 1:00 AM
> To: xri-editors@lists.oasis-open.org
> Subject: [xri-editors] Reviewing RC1b
> 
> Editors,
> 
> I just uploaded RC1b to the Editor's SC for review by the editors before
> posting as RC2 to the main TC (ideally RC2 is final final before the
> vote, so I want us all to review it before posting). For ease of review,
> EVERY change (edit or addition) is commented, so you can just open the
> comment window and click each comment to review.
> 
> Here are the main thrusts:
> 
> 1) Eagle-eye Nat did a great job of: a) providing the example of ASCII
> char sequences that are non-legal UTF-8, and b) proofing the spec (he
> caught the missing dots in some of the resolution examples).
> 
> 2) In discussing the $! metadata issue with me on Friday, Gabe brought
> up the issue of "canonicalization" of XRIs, and we both realized that
> this was a better way to deal with the whole issue of what we had been
> calling "optional syntax". So I added a definition of canonical form,
> changed the title of that section with 2.4, and reordered it to come
> first.
> 
> 3) After Gabe's and my conversation about $!, I added a section called
> "Encoding Presentation Metadata" that is patterned after the "Encoding
> Non-UTF-8 Language and Font Metadata" section. Note that I added this
> section BEFORE the "Transforming XRIs into IRIs and URIs" section
> because it deals with a higher-level concept, namely "Transforming
> Display Strings into XRIs". In other words, the purpose of $! is
> capturing display metadata that otherwise may be lost when converting a
> display string into a valid XRI in escaped normal form. This metadata,
> which can address difficult issues we've been discussing like spaces,
> case, and punctuation, can be lost DURING the transformation into
> escaped normal form depending on an authority's XRI normalization rules
> (which might, for example, disallow even %20 encoded spaces or capital
> letter). So my reasoning was that if this data needed to be retained for
> later display, it needed to be captured into a $! cross-reference at
> that juncture.
> 
> (Note to Gabe: I realized I couldn't add $f and $l to the $! space
> because they ARE significant in comparision, while $! metadata is not.
> That's another reason I added this section prior to the "Transforming
> XRIs into IRIs and URIs" section.)
> 
> Please review all these edits, especially this new section, and post any
> feedback. To my knowledge, this is the only remaining ISSUE in RC1b.
> With Nat's outstanding tasks done, everything else is strictly a matter
> of reviewing new or added wording. So if we can close this issue by COB
> TOMORROW, MONDAY, then I will post RC2 (incorporating any final feedback
> on these edits and any other feedback) Tuesday morning.
> 
> Note that I will be offline from about 9am - 4pm tomorrow, but during
> that time you can get me on my cell, 206.618.8530.
> 
> =Drummond
> 
> To unsubscribe from this mailing list (and be removed from the roster of
> the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/xri-
> editors/members/leave_workgroup.php.
>