Re: AuthConfReq: Presentational Markup

On 03/27/2010 02:25 PM, Tab Atkins Jr. wrote:
> On Sat, Mar 27, 2010 at 10:58 AM, Sam Ruby<rubys@intertwingly.net>  wrote:
>>>> What new mime type do you propose for this?
>>
>> I note that this question was not answered.
>
> I thought it subsumed under my following answer; I don't believe there
> is any need to mint a new mime type.  That would be explicitly
> versioning the web, which I don't think is necessary at this time.
>
>>> So we both agree that the web is unversioned.  We just, apparently,
>>> disagree on quite what that means.  To me, that means that we are
>>> required to support old 'versions' by ensuring that they act/look the
>>> way they were intended, as much as is possible/practical.  It does not
>>> mean (to me) that we have to continue to bless every aspect of every
>>> old 'version' as being equally valid and correct.  Some things were
>>> simply mistakes.  Widely-used mistakes still have to be supported, but
>>> we can still say "Hey, this was a mistake.  Don't use it anymore.".
>>
>> I note in passing the use of the word 'bless'.
>
> I'm not sure what you are trying to imply here.  Could you try to
> state your points less cryptically?

I was rightly criticized for referring to such beliefs as dogma.  I was 
noting in passing a word which has a specific religious connotation:

http://www.google.com/dictionary?langpair=en|en&q=bless

I do not mean to belabor this point, I truly only meant to note this in 
passing.

>>> Doing anything else is abdication of our duty to move the web forward.
>>
>> A wee bit of hyperbole.
>
> I don't think so.  We have the duty to move the web forward.  Saying
> that everything that has ever been done in the past is valid is
> abdication of that duty.  I mean that very literally.  Part of moving
> the web forward is solving new problems.  Another part is identifying
> failed solutions and discouraging their use.

Discourage?  That we can agree on.  Doing anything else than exactly the 
course you have specified is abdication?  That's hyperbole.

>>>   We should help authors, for precisely the reasons listed in the
>>> author conformance requirements,
>>
>> Agreed.
>>
>>> and that means forbidding some constructs
>>
>> Disagreed.
>>
>>> while still specifying how they should be interpreted when
>>> they are encountered.  That does mean that some documents which were
>>> previously conforming aren't anymore.  And?  The page still works.
>>> It's just not authored according to current best practices, /which is
>>> precisely the point/.  Conformance reflects reality.
>>
>> I would have no problem with a Best Current Practices document, such as the
>> IETF often produce[1].
>
> So you think that we can help authors more by explicitly making lots
> of things valid, and then producing a separate document that says not
> to use some of them?  (Not rhetorical - this just seems to be what
> you're saying and I want to make sure.)

The second document wouldn't say "don't use", it would say that under a 
given set of circumstances, there are better alternatives that we suggest.

>> I also don't believe that your notion of conformance reflects reality. The
>> reality I see is that the overwhelming majority of pages violate
>> conformance.  I furthermore don't believe that you can segregate pages into
>> "legacy" and "new".  I believe that there are a lot of not-new but actively
>> maintained and living documents out there.  Many of which willfully violate
>> these conformance requirements.
>
> Sure, most documents are nonconforming.  Most documents aren't
> authored according to what we currently call "best practice".  I'm not
> sure what the problem with this is.  Can you elaborate on what you
> believe the problem is with calling these documents nonconforming?

I believe that we can come to agreement in finite time on what the font 
tag means, how it is to be serialized, and how it is to be processed.  I 
believe that, collectively between the WHATWG and W3C, we have assembled 
the right group of people to do exactly that.

I don't believe that we can come to agreement in finite time on what the 
best current practices are.  I don't believe that we have gathered 
together the right group of people to tackle that task.  I also believe 
that best current practices evolve at a different pace than Standards do.

> On a slightly different tack, if we go your suggested route of making
> all of these valid in HTML5 and then producing a separate Best
> Practices document, the documents that HTML5 currently calls
> nonconforming will instead be conforming HTML5 but nonconforming to
> Best Practices.  Do you believe that this produces a superior state of
> affairs?  If so, why do you think it is okay to say something doesn't
> conform to a Best Practices document, but not okay to say something
> doesn't conform to HTML5 for the same reasons?

You've dropped the word 'Current' which I believe to be an important 
word.  BCPs evolve at a different pace than Standards.  There may be 
valid reasons to chose to not conform to a Best Current Practice.  There 
may be separate Best Current Practices for semantic markup and Best 
Current Practices for accessible markup.

I also believe that discourage doesn't necessarily mean disallow.

If you don't mind, I've answered a lot of questions.  Can you (or 
anybody) explain why discourage must mean disallow?  Or that Best 
Current Practices must be standardized at the same time as the simple 
definition of what these elements mean, how they should be serialized, 
and how those serializations are to be processed?

I think we will get through this process a lot faster if we don't try to 
push forward a registration of text/html which makes the overwhelming 
majority of the web non-conformant.

I'm sure you've seen the following by now, but I include these links for 
others:

http://intertwingly.net/blog/2010/03/20/Authoring-Conformance-Requirements
http://www.w3.org/html/wg/wiki/HTML5_Authoring_Conformance_Study
http://intertwingly.net/stories/2010/03/21/summary.html

If we can get a document which accurately describes the web as it 
exists, I think that would be a huge step forward.  Such a web is messy 
and imperfect.  It contains many aspects which arguably misguided people 
are intentionally violating the best current practices.  It also 
contains a number of aspects where well intentioned people are 
unintentionally making egregious errors of unmatched quotes or inserting 
paragraph elements as immediate child nodes of tables.

The latter is a conformance error: it interferes with our ability to 
understand the meaning of the document.  The fact that HTML5 defines 
consistent behavior in the face of such nonsense is outstanding.

The former, however isn't a conformance error.  It is merely something 
that we currently feel should be discouraged.

> ~TJ

- Sam Ruby

Received on Saturday, 27 March 2010 19:14:45 UTC