Standards compliant webdesign

Hi list

We are probably all in favor of standards compliant webdesign, and we
have just had an intermezzo in the other thread about how to serve pdf
files to the user.
 
Some members of the list raised the issue that one should only use
standard HTTP headers, and that I should not use the Content-Disposition
Header since it is not yet a standard only a *proposed* standard.

I have asked the responsible editor of the IETF, Keith Moore, about what
to do with this *proposed* RFC.  His answer is very interesting, and it
gives us an excellent opportunity to look into the inner workings of a
standards organization.

At least we now know that being "standards compliant" is a very
difficult animal to deal with in the practical world of web design
making web pages.

I quote the full correspondence. Keith Moore has given my permission to
do so:

_ _ _ _ _ _ _ _ _ _


From: Keith Moore
Date: July 29, 2004, 14:57
To: Jesper Tverskov
Subject: Re: Content-Disposition Header

This is a known problem with the IETF process.  lots of specifications 
languish at "proposed standard".  it takes a lot of work to move the 
specification to "draft standard", which is the (confusingly named) 
next stage - you have to do interoperability tests and fix any bugs in 
the specification, which often opens the door to re-editing the entire 
specification (not to make major changes, but to clean up muddy text).  
  When you re-edit the specification, you find that references need to 
be updated because those have advanced, and you may find that the 
document you're trying to advance depends on references that aren't 
advanced yet, which delays the whole process.

IETF is a "volunteer" organization (meaning that people either put 
their own time into IETF activities or - more often - companies pay 
their employees to work in IETF toward common industry goals), and the 
reason volunteers put energy into IETF is to produce specifications.  
Once the specification exists and has been agreed to, there's little 
incentive to advance it along the standards track - unless it's found 
to be buggy or ambiguous.  Thus many of the specifications that advance 
in grade are those which were problematic when first released.  
Content-disposition seems to work pretty well, so there hasn't been 
much interest in advancing it.

IETF is currently discussing how it might change its standard maturity 
levels, to address this and other problems.

And yes, it is safe to use Content-Disposition in accordance with RFC 
2183.  Note that RFC 2231 added a new way to encode (long or non-ASCII) 
parameters that might be used by Content-Disposition.  2231 is not as 
widely supported in existing products as 2183.

Keith

On Jul 29, 2004, at 5:28 AM, Jesper Tverskov wrote:

> Hi Keith Moore
>
> Could you tell me please, what happened to:
>
> rfc2183
> The Content-Disposition Header
>
> Why is it still just *proposed*, and not a standard yet, since all 
> browsers seem to support it, and in MS asp.net you can set it by code:

> http://support.microsoft.com/default.aspx?scid=kb;EN-US;q260519
>
> etc.?
>
> Is rfc2183 safe to use, or should we wait until when?
>
> Best regards,
>
> Jesper Tverskov
> www.smackthemouse.com
>
>

Received on Friday, 30 July 2004 05:16:34 UTC