Hi Andrew - POST responses require the OP to return HTTP 200 with a self-submitting form in the body. This causes the browser to display a blank white page as an interstitial before the RP¹s return_to URL is loaded. Given that the RP¹s return_to page will probably take a couple seconds to load (network latency, verifying the assertion, doing db lookups, etc) the blank white page really looks clunky.
There are a couple hacks that can be done to make the POST form look more attractive, but that involves returning even more data in the response body. In contrast, the 302 Redirect/GET response does not display any blank interstitials between the time the response is sent to the browser and the the RP responds. There¹s plenty of UX research that shows the effect of perceived latency where even a fraction of a second increase in latency results in a measurable decrease in user satisfaction. To encourage adoption, it¹s very important that we match or exceed the UX standards that have already been set by the proprietary solutions. Thanks Allen On 1/28/10 6:02 AM, "Andrew Arnott" <[email protected]> wrote: > On Wed, Jan 27, 2010 at 11:21 PM, Allen Tom <[email protected]> wrote: >> POST adds additional latency, and can cause strange warnings and a blank >> interstitial (the self submitting form). > > I agree with all your points, Allen. But can you explain why POST adds > additional latency? It seems like just a word change over the wire. A browser > and server shouldn't (it seems to me) take any longer to process it, except > that it disables caching to some extent, but in this case that's desired.
_______________________________________________ specs mailing list [email protected] http://lists.openid.net/mailman/listinfo/openid-specs
