Re: ACTION-95, ISSUE-65: Plan to publish a new WD of HTML-5

Philip TAYLOR (Ret'd) wrote:
> 
> 
> Jonas Sicking wrote:
> 
>  >> Is the simple answer to this question not just
>  >> "because if it is non-normative, it is of no use" ?
>  >
>  > So would you say that the documents Dan and Lachlan have produced is
>  > of no use since they are informative?
> 
> No, to do so would be insulting.  Rather I would say
> that they are of some use, and that they may well
> be of interest to many seeking to exploit HTML 5,
> but that any informative document can only augment,
> rather than replace, a normative document that addresses
> the same topic (cf. "An Informal Introduction to Algol 68",
> v. "The Revised Report on the Algorithmic Language Algol 68").
> Titles from memory, so don't take me up on typos.

Is the augmentation rather than replacement a problem? If so, why?

Also, why would Mikes document be of no use if it's non-normative, but 
Dans and Lachlans not?

>  > When you write a perl program, do you read the Pod documentation, or
>  > do you go read the perl source code (which as far as I can tell is the
>  > only thing resembling a spec for Perl5-)
> 
> I use "Programming in Perl".  For TeX, of which I write far
> more, I use "The TeXBook", "TeX the Program", Eijkhout's
> "TeX by Topic", and "Tex in Practice - 4 Volumes":
> Stephan v. Bechtolsheim.  For Pascal, the "User Manual
> and Report".  And so on.  In other words, my preference
> is for the most definitive (and, where possible, normative)
> reference on any language in which I program.

You'll have to pardon my ignorance, but is any of those TeX books a 
normative TeX specification?

For Pascal I notice that you don't include any of the Pascal ISO 
specifications? Why is that?

>  > In my experience only experts in a language ever go look at the
>  > specification. They are simply too detailed to give non-experts enough
>  > of a high-level view that the information can be consumed. Non-experts
>  > tend to go to other resources that provides easier-to-consume
>  > information.
> 
> Exactly.  Hence the need to separate the need-to-know part
> (markup) from the "only if you need to process this stuff"
> part, and make the former accessible to all webmasters.

Why is only the markup need-to-know?

http://code.google.com/webstats/2005-12/pages.html

shows that at least 2/3 of all pages on the web in 2006 used scripts, 
(not counting pages that only use on* type attributes). In fact, 
scripting is more common than the majority of markup elements.

> But unless the part factored out is normative, the poor
> webmaster may /still/ have to refer to the full HTML 5
> Specification, and that is what I believe we must avoid.

Ultimately the user will have to refer to a normative recommendation. I 
believe any such document is going to be strictly for expert users due 
to the level of details a normative recommendation has to have. If that 
ends up being a separate recommendation or a separate section in the 
main HTML 5 recommendation I think makes little difference to the user.

/ Jonas

Received on Wednesday, 28 January 2009 23:23:03 UTC