"downplayed errors"

On Thu, 5 Feb 2009, Henri Sivonen wrote:
>
> Recently, though, HTML 5 has gained a list of of downplayed errors[2] 
> which, in my opinion, is partly similarly wishful as deprecation was in 
> HTML 4.01.
> 
> [2] http://www.whatwg.org/specs/web-apps/current-work/#conformance-checkers-0

I agree with that characterisation.

The intent of the downplayed errors is to make migration easier while not 
leading authors of new documents into the weeds. It's an attempt at 
addressing two mostly-contradictory requirements:

 1. Don't waste the time of new authors by having them look at features 
    that are useless or otherwise harmful,

 2. Don't waste the time of authors who are editing existing documents 
    that may contain those mostly pointless features but for which the 
    investment is a sunk cost.

There's an inherent conflict when one is trying to use document 
conformance as a tool to both prevent people from writing new material and 
as a tool to allow people to keep existing cruft untouched.

If anyone has any better proposals for addressing these two use cases 
simultaneously, please do bring them up. Also, if anyone sees the working 
draft slip down the slope of using the "downplayed errors" category as a 
way to avoid answering hard questions, then please do point it out.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Thursday, 5 February 2009 08:10:09 UTC