W3C home > Mailing lists > Public > public-html@w3.org > January 2009

Re: ISSUE-54: doctype-legacy-compat

From: Leif Halvard Silli <lhs@malform.no>
Date: Fri, 09 Jan 2009 02:11:59 +0100
Message-ID: <4966A45F.3020107@malform.no>
CC: public-html@w3.org

Shawn Medero 2009-01-08 23.36:
> So I'm wrong. I still prefer "legacy-compat" to an empty
> sting... but a suggested technical barrier to using the empty
> string might bumpkis.

But why support a doctype based on principles quite different from 
principels behind the choice of <doctype html>?

<doctype html> was chosen as the shortest common browser 
denominator. If some author tools requires another common 
denominator, then the shortest denominator ought to be chosen for 
that usecase as well.

Despite shortness, <doctype html> is seemingly meaninfull (aka 
doctype="html") and unambiguous. In comparison,  <doctype html 
public ""> is dull and meaningless - it would probably mostly go 
unnoticed - authors would not think about it even if they saw it. 
Its dullness is also medicine against overuse, though. Wheras on 
the contrary <doctype html public "legacy-compat"> is open to much 
interpretation and is also an attention grabber. Plus it is more 
open to misspelling.

(Amongst the possible misunderstandings that could arise, I would 
mention that "quirks-mode" has sometimes been translated as 
"compatibility mode" - at least in Norwegian.)

Henri Sivonen has some examples [1] of how authors attribute false
meaning to e.g. the "//EN" fragment in the doctypes in use today. 
Hence I conclude that to have anything seemingly meaningful inside 
the DOCTYPE is bad. Thus I prefer the empty string.

I suspect that many of the supporters of "legacy-compat" think 
that such an negative string will work in favour of the short 
doctype - <doctype html>, which they consider as the "real" HTML 5 
doctype. But I see no basis for such a belief.

Do we want authors to correct themselves if they use the long one? 
Why? Most sane persons and tools would automatically chose the 
shortest doctype - short and meaningfull is good! Hence, to choose 
a string equivalent of "don't use this doctype" for the legacy 
doctype seems counterproductive.  It gives the wrong kind of 
attention to the doctype.

After all, the draft says that the doctype is "mostly useless". If 
so, then it is pointeless to add a negative "don't use this 
doctype" string to the long doctype.

The simplest  - when first two doctypes are needed - is that the 
doctypes are based on the same principles. And many authors wiill 
naturally believe that they *are* based on the same principles. 
Thus, to go for "legacy-compat" can give the impression that there 
is is a hidden meaning (aka doctype="html") even in <doctype html> 
(as I explained above). Wheras to go for the empty string can help 
underline the fact that both doctypes are as short as possible.

OTOH, the apparent meaningfullness of <doctype html>, along with 
its shortness, will anyhow work in favour of the short doctype. 
Which is another reason for not adding a needless and negative 
non-empty attentiongrabbing string for the compatibility docytpe.

[1] http://hsivonen.iki.fi/wannabe
-- 
leif halvard silli
		
Received on Friday, 9 January 2009 01:12:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:28 GMT