Wow @Jay, that's a bit of a lecture there ;)

@Ryan, we've tried and tested the iframe approach you suggested and
I've got my answer:

Yup, you're right. The reason is speed, loading an iframe causes the
main ajax request to take about 30% longer in averege which is a lot
(even if it's only 100-200ms) so I gotta shelf the idea, at least
until I can get the speed facebook has achieved. Good lesson anyway!

Cheers all




On Jul 12, 10:23 am, André Fiedler <[email protected]>
wrote:
> @Jay That´s great! That´s the way i´m doing this, too. ;o)
>
> 2010/7/12 Jay <[email protected]>
>
> > I find my most creative work comes from being restricted from doing
> > something the way I'd initially want to do it. When being forced to
> > work within confined limits, you can often come up with an elegant
> > solution that's cooler and more usable than existing implementations.
> > Instead of having wacky IFrames and stuff, what if you would use a
> > slightly-less-standard paradigm for user feedback? Start with the
> > basics. To preach to the choir, I almost ALWAYS set the cursor
> > property if I'm doing time-consuming requests, as Mr. Newton
> > suggested. I'm not sure why, but even something as subtle as that
> > seems to have a huge effect on the user of the site. If you think you
> > need more feedback than that, what if you did a "Loading..." message
> > in the style of GMail? Something like that could integrate beautifully
> > with the design of your site, and would give your app a really novel
> > feel.
>
> > But, of course, it always depends what you're going for. If you really
> > really really need to simulate that browser behavior, IFrames are the
> > only way to go. However, I imagine that the extra requests to IFrames
> > + extra load on your server + JavaScript overhead would make your AJAX
> > site load just as slowly (if not more so) than a traditional site,
> > which seems to defeat the purpose of AJAX in the first place
> > (especially when you consider its usability concerns for the outskirts
> > of the internet world -- screen readers, old web browsers, javascript-
> > blocking firewalls and paranoid javascript disablers. Coincidentally,
> > the people I find I need to emulate "traditional" methods for to avoid
> > confusing them are the same people that -- of course -- use screen
> > readers, old web browsers, and the like. At which point, I make sure
> > the site gracefully degrades, and kill two birds with one stone.
>
> > On Jul 10, 4:28 pm, Ryan Florence <[email protected]> wrote:
> > > :D
>
> > > Momentary lapse of reason...
>
> > > Sent from my iPhone
>
> > > On Jul 10, 2010, at 10:35 AM, jiggliemon <[email protected]> wrote:
>
> > > > @Ryan
>
> > > > OMG <--- meant to be rude.
>
> > > > -Chase
>
> > > > On Jul 9, 9:37 pm, Aaron Newton <[email protected]> wrote:
> > > >> <facepalm>
>
> > > >> On Fri, Jul 9, 2010 at 2:50 PM, Ryan Florence <[email protected]>
> > wrote:
>
> > > >>>> For now I'm thinking about a class, similar to Request.HTML that
> > gets
> > > >>>> stuff into an iframe, loads it, then returns the content into
> > > >>>> onSuccess and removes the iFrame? Not AJAX, but might work - would
> > > >>>> that make sense?
>
> > > >>> If I  was forced to code it I would have the request be a normal
> > request
> > > >>> and then add onRequest and onComplete events.
>
> > > >>> onRequest I'd create an iframe element with a src pointing to a file,
> > that,
> > > >>> on the server side (assuming php) is just a big long sleep(), like
> > > >>> sleep(10000000000).  That should force the browser into a "loading
> > state".
>
> > > >>> And then onComplete destroy the iframe element.
>
> > > >>> I love to beat dead horses.  A simple indicator by the link or over
> > the
> > > >>> content that's updating is enough ...

Reply via email to