Re: [vip] Draft on IDN Tables in XML
Hi Thomas, On Mar 1, 2012, at 12:21 PM, Thomas Roessler wrote: A few quick comments on the XML format level upon skimming this (mostly nit-picking): - you don't seem to identify the namespace URI in the normative text, it's hidden in the RNG schema - you don't use the namespace in your examples. - the example in 3.4 doesn't follow the specification for the description element: You either need a CDATA construct, or you need to entity-escape the pointy brackets. - in the specification for the description element, don't talk of "CDATA" as an element. It isn't, but rather a particular way to mark up textual data: http://www.w3.org/TR/xml/#sec-cdata-sect So: The content model for <description> is text; if content includes markup, then CDATA is a convenient method to write that out. - MIME-type -> media type All great fixes I'll work into the next revision. - To the unique identifier point in your issues list -- agree that it might be useful to give the table a unique identifier. I probably wouldn't choose an ID attribute, but would include a "this version" and "latest version" URI in the <meta> section. Sounds good to me. - I may be missing something, but what's the extension model for the @when attribute on <var>? Are certain names there reserved for IANA, and others up for grabs? Want to use URIs for these names, so everybody can coin their own without stepping on others' toes? Something else? Good points, the mechanism for specifying "when" attributes is undefined right now. Right now I know to implement some tables we need to support context in Arabic script strings. I don't know if it is knowable in advance all the types of contextual modes that will need to be supported. If we assume they can not be, then I agree there needs to be a mechanism to keep track of them so the format doesn't extend in competing directions and lose its utility. Hope these are useful, Very, thanks! kim
http://iucg.org/wiki/IUWW_Draft-Davis-IDNTables There is a proposition of Draft by Kim Davies (ICANN) for IDN Tables in XML format. The discussion is reproduced on the talk page. This work seems interesting and could be used as a root model to extend for "rootlocale" files in the MDRS documenting the relational space of an Internet TLD. There is however two missing metadata: - the encoding scheme (unless we accept Unicode as a default value) - the CLASS (unless we accept the ICANN/NTIA CLASS "IN" as a defaulf value) jfc
Greetings! What about "one-direction" variants? For example, Russian letter ё can be replaced with е and often is, but the letter е can not be replaced with ё. Thank you! On Fri, 2 Mar 2012, JFC Morfin wrote:
http://iucg.org/wiki/IUWW_Draft-Davis-IDNTables
There is a proposition of Draft by Kim Davies (ICANN) for IDN Tables in XML format. The discussion is reproduced on the talk page. This work seems interesting and could be used as a root model to extend for "rootlocale" files in the MDRS documenting the relational space of an Internet TLD.
There is however two missing metadata:
- the encoding scheme (unless we accept Unicode as a default value) - the CLASS (unless we accept the ICANN/NTIA CLASS "IN" as a defaulf value)
jfc
-- SY, Dmitry Belyavsky
On Fri, Mar 02, 2012 at 02:14:59PM +0100, JFC Morfin wrote:
- the encoding scheme (unless we accept Unicode as a default value)
You had another alternative? IDNA uses Unicode. Adding an alternative encoding is just begging for trouble.
- the CLASS (unless we accept the ICANN/NTIA CLASS "IN" as a defaulf value)
Why does this need to be classed? I can see a reason to add it: maybe a future CLASS would not need any ACE. But thought experiments, proposals, and experience with other protocols so far suggest that it would be easier to create DNSng than it would be to add a new widely-supported CLASS to the DNS. We can't even get new RRTYPEs widely supported in the DNS. A -- Andrew Sullivan ajs@anvilwalrusden.com
On 7 mar 2012, at 17:39, Andrew Sullivan wrote:
On Fri, Mar 02, 2012 at 02:14:59PM +0100, JFC Morfin wrote:
- the encoding scheme (unless we accept Unicode as a default value)
You had another alternative? IDNA uses Unicode. Adding an alternative encoding is just begging for trouble.
"utf-8" as mentioned in the email header is not the name of an encoding. It is a charset as registered with IANA. Patrik
participants (5)
-
Andrew Sullivan -
Dmitry Belyavsky -
JFC Morfin -
Kim Davies -
Patrik Fältström