Re: References Re: What are the requirements/problems? Re: Working on New Styles for W3C Specifications

On Wednesday, 14 December 2011 at 22:14, fantasai wrote:

> On 12/13/2011 07:34 AM, Harald Alvestrand wrote:
> > =
> > The references need to show what the spec author was referring to when
> > he wrote the text he wrote. If a spec is updated after that, either
> > the reference will still be valid (no problem), or it will be broken
> > because of the update - in that case, the people reading the spec need
> > to know that the text they are reading does NOT refer to the newer
> > specification.
> 
> 
> 
> I've actually run into this problem with references to UAX29.
> Older versions just defined "grapheme clusters". Newer ones
> have "legacy grapheme clusters" and "extended grapheme clusters".
> If there wasn't a dated reference, it would be unclear what the
> editor meant. So I like Karl's suggestion to include both.

I agree that citing both could work. Reminds me that at university we always had to cite the "date accessed" of a Web resource with the URL. So, you could basically keep pointing to the latest version, but also include a link to the cited dated version as fallback in case things go bad (such as defined concepts changing name on you). 

-- 
Marcos Caceres

http://datadriven.com.au 

Received on Wednesday, 14 December 2011 22:27:50 UTC