- From: <noah_mendelsohn@us.ibm.com>
- Date: Fri, 9 Mar 2007 14:17:19 -0500
- To: Jon Hanna <jon@hackcraft.net>
- Cc: www-tag@w3.org
I think the point is that doing a GET is not supposed to incur any obligations. Rfc 2616 says: "the convention has been established that the GET and HEAD methods SHOULD NOT have the significance of taking an action other than retrieval. These methods ought to be considered "safe" ... Naturally, it is not possible to ensure that the server does not generate side-effects as a result of performing a GET request; in fact, some dynamic resources consider that a feature. The important distinction here is that the user did not request the side-effects, so therefore cannot be held accountable for them." So, proxies are one example, which I've given, so are mechanisms that pre-cache access. If some user agent running on my system chooses to dereference a URI, that might leave an audit trail suggesting that "I" have accessed the information. Whether that's what you want will depend on the circumstance, but I claim there are circumstances where that's not what you want, and that you shouldn't use GET in such cases. I do agree that leaving a mark in an audit trail is a less clear cut issue than, say, cancelling a subscription when someone clicks on a link. Some audit trails are used primarily for things like debugging, and clearly it's reasonable for that side effect of the GET to occur. Given legislation like Sarbanes-Oxley in the United States, an entry in an audit trail can conceivably have legal implications at least as great as cancelling a magazine subscription. In such cases, I think GET is inappropriate, and I'm not convinced that must-revalidate headers in general solve all the problems. Noah [1] http://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html#sec9.3 -------------------------------------- Noah Mendelsohn IBM Corporation One Rogers Street Cambridge, MA 02142 1-617-693-4036 -------------------------------------- Jon Hanna <jon@hackcraft.net> Sent by: www-tag-request@w3.org 03/06/2007 07:14 AM To: www-tag@w3.org cc: (bcc: Noah Mendelsohn/Cambridge/IBM) Subject: Re: State, when to use GET, and accountability noah_mendelsohn@us.ibm.com wrote: > I'm not concerned so much with cases in which a cache is being used on > behalf of a single user of HTTP; my real concern was about shared proxy > caches, which are quite common. With those, you might request some > information using HTTP GET, and your access would be appropriately > recorded at the origin server. My concern is with the case in which I > come along to request the same information, but a proxy cache somewhere in > the network satisfies the request without contacting the origin server. My > access is likely not recorded, except in the unusual special case that the > proxy and the origin server are cooperating on maintaining the audit > trail. Is this really an issue. Sure someone could fail to send must-revalidate headers, but they could also forget to make sure they logging application has write permissions where it should be logging and the logs aren't all silently failing. Bugs happen, but the specs cover this concern already.
Received on Friday, 9 March 2007 19:17:33 UTC