W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2011

Re: URI equivalence and query parameter sequence

From: Dale Anderson <dra@redevised.net>
Date: Wed, 13 Jul 2011 21:55:58 -0700
Message-ID: <CANNRn6+UwS9esQq-QE0k95_tNObiQBy859ckgT3zEtcS5iRerA@mail.gmail.com>
To: Darrel Miller <darrel@tavis.ca>
Cc: ietf-http-wg@w3.org
Darrel,

In terms of how the original web server interprets the request, named query
parameters' order may likely matter only in the case of trump or order when
repeatedly specified query parameters.

But, keeping them consistent may help avoid cache pollution in a simple
caching scheme that does not evaluate the query string meaningfully, rather
keys off the literal URI.

Also bear in mind the case of unnamed query parameters like ?buy&cat which
are certainly important as far as their identity is their index ordinal
position in the series

In my estimation!

Regards,

Dale

2011/7/11 Darrel Miller <darrel@tavis.ca>

> To get to the point, I would like to know if the following two URIs can be
> considered equivalent.
>
>   http://example.org/location?x=34&y=67
>
>    http://example.org/location?y=67&x=34
>
> The HTTP spec refers to RFC 3986 for rules on normalization and comparison.
> In RFC3986, it does discuss the use of "syntax based normalization"
> (http://tools.ietf.org/html/rfc3986#section-6) but makes no mention of
> query
> parameters.
>
> It is my understanding (and quite possibly incorrect) that ordering of
> query
> parameters is not significant to most server implementations.  Is there any
> official opinion on this?  Is it reasonable to consider these two URIs as
> accessing the same resource? Or should I redirect one URI to the other?
>
> Thanks,
>
> Darrel Miller
>
>
>
>
>
Received on Thursday, 14 July 2011 04:56:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:44 GMT