RE: 301/302

Klaus you were absolutely right, I didn't understand the solution and I
apologize.

I was able to catch the bits at the last second, we are that close to
ship. We have made the 307 change as specified but this is it. We can't
change again. So we are shipping with 307.

Either way, however, I think this is a really good solution. It supports
everyone, it doesn't break anything, and best of all - it was easy to
do. =)

Thanks to everyone. It is this sort of quick decision making that allows
us to create open standards but still react to the realities of the
market.

			Yaron

> -----Original Message-----
> From:	Klaus Weide [SMTP:kweide@tezcat.com]
> Sent:	Wednesday, September 03, 1997 10:05 AM
> To:	Larry Masinter
> Cc:	Yaron Goland; http-wg@cuckoo.hpl.hp.com
> Subject:	Re: 301/302
> 
> On Tue, 2 Sep 1997, Larry Masinter wrote:
> 
> > We seem to be going around in circles on this one.
> > 
> > Yes, some people prefer the proposal expanded in the message
> >    http://www.ics.uci.edu/pub/ietf/http/hypermail/1997q3/0402.html
> > 
> > which isn't just "add 307", but a complicated "add 307,
> > deprecate 302 but leave it there anyway".
> 
> You have just shown that it can be nicely summarized in a few words.
> 
> > Most people at the meeting in Munich seemed to prefer the
> > alternative, which is just "swap 302 and 303".
> > 
> > We know there is some argument for cleanliness, for those who
> > went ahead and implemented 303 as it was originally described.
> > But it was felt "the cat's not really out of the bag". So if
> > it isn't, can we just go ahead and do the swap?
> 
> So if it _is_, can we just go ahead and do the "add 307, deprecate
> 302 but leave it there"?  
> 
> In my view the 302 cat was out of the bag when 1945 was published, 
> and the 303 cat out of the bag when 2068 was published.  RFCs _do_ get
> read by some people, 303 was discussed on newsgroups and appears in a
> CGI FAQ, it is likely that it appears in some books on the topic. Now
> the
> choice is between making completely (and unnecessarily) incompatible
> protocol changes on the one hand, and changing only what contradicts
> with
> reality OTOH.
> 
> > We need to decide this one soon.
> 
> Is this is more urgent than other ISSUEs?
> 
> Either way, for the main practical question which triggered this
> issue,
> whether issuing GET in response to a 302 response to a POST is
> allowed,
> the answer will be the same: yes it is.
> 
>    Klaus

Received on Wednesday, 3 September 1997 14:02:44 UTC