ietf-822
[Top] [All Lists]

Re: X-* header fields

2004-01-23 04:31:18

At 00:43 23/01/2004, Dave Crocker wrote:
My own, primary concern about them is having an "experimental" header
become a defacto standard and then having to find a way to get folks to
use a new, non-X header name for it.  I believe that hasn't been
successful yet, or at least not much.

I think it is arguable that you need some method of having 'experimental' headers 'in the wild'.

Getting rid of X- headers and saying 'you must come up with an agreed standard before you can add a new header of any sort' is going to inhibit innovation.

At least if you use X- headers, and they become defacto standards, then you can tell people off ;-)

One example I can think of of standards changing and people keeping up is with SMTP authentication. In the initial drafts of SMTP authentication the server would say 'AUTH=LOGIN'. In the final standard, the server says 'AUTH LOGIN'. Lots of people used the initial draft because authentication was needed so much, but then changed it to AUTH LOGIN (or supported both, in many cases) when the final standard came out.

I don't see a big problem with that. Yes, it would have been nice to get it right the first time, but if people hadn't been willing to risk getting it wrong and then having to put it right, we might never have ended up with an SMTP authentication standard!

If you look at other situations, eg the extra trace information that people are discussing in the SMTP groups, it's quite possible that someone will come up with a workable idea for this and then people may implement it, and later problems may be found, and the information will have to change.

If that's done, then either we'd have to use a header like 'Originator', and then risk total inoperability by changing the meaning of that header, or make 'OriginatorV2', 'OriginatorV3' etc when the changes happen, and not be able to use the nice 'Originator' ever again, because that's already been used. Or, we could use 'X-Originator', then 'X-Originator2' etc, but then, when everyone's happy with the idea, we could use 'Originator', and deprecate the 'X-Originator' ones. Over time people would move to the Originator header (they may still READ the X-Originator ones, but wouldn't write them)

(This is just an example - yes, it would be nice to get it right first time, but that doesn't always happen cf SMTP authentication).


Paul                            VPOP3 - Internet Email Server/Gateway
support(_at_)pscs(_dot_)co(_dot_)uk                     http://www.pscs.co.uk/



<Prev in Thread] Current Thread [Next in Thread>