- From: Ian Hickson <ian@hixie.ch>
- Date: Mon, 13 Dec 2004 19:30:04 +0000 (UTC)
On Sat, 11 Dec 2004, Jim Ley wrote: > > bittorrent or RSS isn't relevant to the mobile phone user, or the host > of websites, both of these care about the bandwidth of what they're > actually using, reducing bandwidth in requests makes the sites more > usable, as the request size is smaller. Oh, I assure you RSS is relevant to the mobile market. My point was that cutting down on bandwidth by removing fields from the submission does almost nothing to help anyone as far as bandwidth goes. The form submission doesn't even appear on the radar of anything, at this point. > > The processing power doesn't seem like a big thing either -- comparing > > strings is not that expensive, and you have to do it anyway, to test > > for data integrity and prevent invalid data from being submitted, etc. > > No you don't, if there's no FOO field submitted, you don't need to do > any comparison on it for anything, it's just immediately rejected. You have to check it is given or not. The extra cost of checking what value it has is minimal. > > Not to mention that you'd have to do the comparisons anyway to support > > down-level clients. > > There's lots of features in WF2 that are only relevant to new clients, > so what's the problem with this one? It doesn't give any advantages (not any particularly important ones, at least). The other features, like the repetition model, which require server-side support for both models, at least provide important advantages, like not requiring server round-trips for each click. Another problem with not sending fields that haven't changed is that it wouldn't be compatible with checkboxes and multi-select controls (they would act the same as those that haven't changed). -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 13 December 2004 11:30:04 UTC