Posted by Mitja on 06/18/05 21:45

On Sat, 18 Jun 2005 20:03:02 +0200, Joel Shepherd <joelshep@ix.netcom.com>
wrote:

> In article <op.sskd3ipqdob4us@ibis>, Mitja <nun@example.com> wrote:
>
>> the only thing at which POST is better is sending large chunks of data
> No. POST and GET have entirely different semantics, beyond how the query
> data is sent to the server.
>
> http://www.cs.tut.fi/~jkorpela/forms/methods.html

True, I missed that - the implied semantics are indeed quite different.
But - see below. I was perhaps a bit too practical and thinking too much
along the lines of OP's specific case:

> If submitting the form is expected to result in some important change of
> state on the server side (e.g., an order being placed, form data being
> saved in a database, someone's pager going off), then POST should be
> used.
OP's description hardly satisfies this criteria.

> In addition, responses to "GET" requests can be cached (by the browser,
> by a proxy, etc.), while "POST" responses generally are not.
True, and I feel this is the main point I should have mentioned. It can,
however, be circumvented by using appropriate headers. And, honestly -
would you ever implement a plain-text link as a combination of POST and an
invisible form just to try enforcing a clean reload? I still feel GET is
more appropriate in the given scenario (if I got its description right).

[Back to original message]