Re: Support for Design Principles (was: Design Principles survey delayed)

On 15 Aug 2007, at 19:49, Dan Connolly wrote:

> I have an offer from Geoffrey Sneddon to summarize the level of
> support and dissent for most of the principles. Perhaps that
> will be available for discussion in tomorrow's teleconference.

This is what I believe is the general consensus having read most  
emails on this mailing list since it began. If anyone wants to  
question any of these, feel free to do so. This corresponds to the  
current version on the wiki, which due to various changes has a  
greater degree of support.

Support Existing Content: The vast majority seem in favour of this,  
though one or two are trying to avoid deviating from SGML (this is  
perhaps against the charter's requirement to "define conformance and  
parsing requirements for 'classic HTML'", though, making many of  
these really objections against what we've been chartered to create).

Degrade Gracefully: I've not to my knowledge heard anything against  
this, though there have been people questioning how it has been  
implemented.

Don't Reinvent The Wheel: There's has been agreement on this since it  
was reworded by Lachlan Hunt (20070813), though one or two minor  
editorial points have been raised. It seems most of the former issues  
were due to how it was worded and misconceptions about what it was  
meant to be.

Evolution Not Revolution: Very little has been said about this at  
all, but no objections.

Solve Real Problems: This has caused some of discussion about what a  
"real-world problem" is. This is another principle that I believe  
would get the vast majority supporting it if it was better explained  
(such as what a "real-world problem" is).

Priority of Constituencies: This is another principle that has been  
mentioned very rarely, without objections. It was brought up recently  
in terms of accessibility and the needs of users.

Media Independence: Again, been mentioned very rarely without  
objections (mainly by Maciej, and since mid-May, only by Maciej).  
There is a recommendation to add a paragraph in the wiki (seemingly  
the only reason for it to appear under the disputed section).

Support World Languages: This has unanimous agreement, as far as I  
can see.

Secure By Design: This, too, has had very little said about it. I  
doubt anyone would question it, however.

Baby Steps: Having been added very recently, most of the discussion  
under a thread of a similar name ("Baby Steps or Backwards Steps")  
have been regarding the recent changes to @alt, and not to the design  
principles. This should probably be merged into "Solve Real  
Problems", to fix the issues there.

Well-Defined Behavior: There is agreement upon this, though some have  
questioned whether this should spread as far as non-conforming  
content (see "Handle Errors" below).

Avoid Needless Complexity: This has scarcely been discussed, and  
consensus can't really be commented on.

Handle Errors: Some have questioned whether our specification that  
defines parsing requirements should go as far as dealing with non- 
conforming content or not, however, the majority are in favour of this.

Accessibility / Universal Access: The only objection I'm aware of is  
the one directly inline i the wiki, questioning the use of "when  
possible". As this is the only objection, there is a vast majority in  
favour of this.

No Version Syntax vs Explicit Versioning: This is present verbatim in  
the current editor's draft. I assume this is questioning a prior  
principle that doesn't exist anymore?

Pave The Cowpaths: Lachlan Hunt's email of 20070815 <http:// 
lists.w3.org/Archives/Public/public-html/2007Aug/0472.html> explains  
that the principles are there to guide research, not to make  
decisions. If this was clarified, I expect the objections (such as  
condoning past bad practices) would go. This is as it stands, disputed.

Separation Of Concerns: This is disputed on grounds of "semantics for  
the sake of semantics" being good, and not being purely semantic.

Support Extensions: This is something that's come up several times,  
but the technical feasibility of this has sometimes been questioned.  
This probably needs expanding: Is this meaning extensions through  
namespaces, or just putting further elements (such as <canvas>)  
within the global namespace? The WG is currently split over this.  
Such an explanation may help.

Visible Metadata vs. Metadata Anywhere: There has been questions  
about whether metadata should be visible or not. The WG is very split  
over this. This is unchanged from the editor's draft.

It does, in my opinion, appear as if the majority of the document has  
the majority supporting it. I believe by rewriting parts we can  
increase the size of this majority. However, I do believe we may well  
would in a survey get dissent from the WG. However, in such a large  
WG I doubt we'll have many decisions that resolve without dissent,  
due to the huge number of different points of view, so we will in  
many cases just need to manage dissent to make progress (in line with  
the W3C Process). To advance a TR (in this case, to FPWD), the  
Director reviews any Formal Objections.

- Geoffrey Sneddon

Received on Thursday, 16 August 2007 10:54:57 UTC