Re: Response to unknown query parameter

Xiashu Wang writes:

> I found the common practice of web server is to ignore the unknown query 

> parameter.
> 
> For instance,
> 
> http://www.w3.org/?foo=bar
> 
> response the same as
> 
> http://www.w3.org
> 
> But should it be a 404 instead?

Speaking for myself, and not for the TAG,  when you say "should it be a 
404", I think that can be read two ways:

1) Is it strictly a violation of Web Architecture?

I think the answer to that is no.  The authority responsible for these 
reasources is presumably indicating that both http://www.w3.org and 
http://www.w3.org?foo=bar (I'm going to leave out that trailing / in your 
second one) are URIs associated with resources.  Implicit in your example 
is that those resources seem in all cases to serve the same 
representations, and so in the absence of other information external 
observers may at least suspect that they are in fact the same resource.

2) Is it a good idea?

Well, Web Architecture suggests [1] as good practice:

"A URI owner SHOULD NOT associate arbitrarily different URIs with the same 
resource."

So, the question is whether these differences are sufficiently arbitrary 
as to be discourged.  My personal feeling is: in most cases, yes the 
advice in Web Architecture applies here, so don't do it.  Still, this is 
just a good practice, and not strictly required by the architeture;  if 
you feel that the benefits to your users and the community of giving the 
multiple URIs to the same resource outweighs the drawbacks then you can do 
it.

In any case, my personal feeling is that this is already pretty well 
covered by the Web Architecture document.  I suppose that your examples 
might be considered for inclusion in an expanded version of that section 
if the document is ever revised, or perhaps they might be included if we 
decide to update the "Metadata in URI" finding [2].  Thank you very much 
for your comment.

Noah

[1] http://www.w3.org/TR/webarch/#uri-aliases
[2] http://www.w3.org/2001/tag/doc/metaDataInURI-31.html

--------------------------------------
Noah Mendelsohn 
IBM Corporation
One Rogers Street
Cambridge, MA 02142
1-617-693-4036
--------------------------------------








Xiaoshu Wang <wangxiao@musc.edu>
Sent by: www-tag-request@w3.org
08/28/2007 06:49 AM
Please respond to wangxiao
 
        To:     "'www-tag'" <www-tag@w3.org>
        cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
        Subject:        Response to unknown query parameter



Hi, Tag members,

I found the common practice of web server is to ignore the unknown query 
parameter.

For instance,

http://www.w3.org/?foo=bar

response the same as

http://www.w3.org

But should it be a 404 instead?

I think if this behavior is not specified, it will have serious 
consequence because one resource can in fact be identified by many 
URIs.  Should a recommendation be made, or there is already one that I 
am not aware of, regarding the response to unknown query parameter?

Xiaoshu

Received on Tuesday, 28 August 2007 14:18:16 UTC