Re: process and editing questions: RFC errata

hello tom.

On 2013-02-21 17:40 , t.petch wrote:
> In passing, some lines of formal definition exceed the permitted length
> for an RFC line so a little reorganising will be needed, probably best
> sooner rather than later as they will need validating before the I-D can
> advance and reorganising can introduce syntax errors.

thanks for noting, 
https://raw.github.com/dret/I-D/master/xml-patch/draft-wilde-xml-patch-05.txt 
should look better (not yet submitted).

> I may have missed the errata but I cannot recall a reference to them on
> the apps-discuss list.  If they are modified and then approved, which
> usually happens, then your I-D should follow suit so the sooner they are
> processed the better.  Which might mean you requesting the AD to set the
> wheels in motion, explaining why timeliness matters.

the errata are still just in the "reported" state, 
http://www.rfc-editor.org/errata_search.php?rfc=5261 lists the 4 i have 
submitted. 3 of those now are actually part of the updates to RFC 5261 
in the draft, so i am wondering whether these errata are needed anymore? 
ideally, my draft would update RFC 5261, and then the errata would be 
redundant, right?

> And, out of curiosity, do you expect people to use XPath 1.0 or 2.0?  I
> ask because in Netconf, I was keen to specify 2.0 and not 1.0, the
> handling of namespaces in 2.0 seemed superior, but was told we could not
> because there were no implementations for people to use, that 2.0 was a
> great idea that had not happened (mmm IPv6?).  The Normative reference
> for Yang remains the 1999 version.

2.0 is a vast improvement over 1.0, but it also is much more complex to 
implement. when 1.0 was released, XML was all the rage and there were a 
lot of people implementing specs. when 2.0 was released, the XML hyper 
curve already trended downward, plus it's just harder to implement. as a 
result it's true that it's surprisingly hard to find implementations of 
2.0. so while personally, i always use 2.0 because you can write better 
code, it's true that in specs, if you can get away with 1.0, it may be a 
good idea to stick to it.

cheers,

dret.

-- 
erik wilde | mailto:[email protected]  -  tel:+1-510-2061079 |
            | UC Berkeley  -  School of Information (ISchool) |
            | http://dret.net/netdret http://twitter.com/dret |