Just in case this is someone I know (as I am one of a group of guys who play Magic - and Munchkin and Dominion and Settlers of Cataan - at lunchtime at work), I do understand this story :)

Appreciated, but there are two reasons we wanted the object/embed code kept:

A) future-proofing

B) a screenshot is just that, a screenshot. A picture on the page, with no context. If the screenshot is loaded by the object/embed code, the screenshot shows not only a "Play" button and QT logo on the iPhone, but the exact same code can be used for other platforms without changes, making maintenance far easier.

We do use SWFObject for publishing the Flash player, but as noted, there is no Flash on the iPhonee, so no, it won't work there.

Uh, actually you CAN browse YT from Safari on the iPhone, and it DOES play videos, and I KNOW there's no Flash on the iPhone (I even said so in the original post).

Prick. Talk about not doing any research...

Incidentally, tell me a better way to play video through a web page than Flash and I'll listen, despite your clear lack of social and relationship skills.

So you know, I wasn't asking about the codec - YouTube seems to have a smooth way of presenting itself to the iPhone and I was looking to replicate it (since YT doesn't publish it's code, obviously). All of our videos are already in h.264 format for use with JW-FLVPlayer (won't work on iPhone) and with the exception of one clients assets, they're all 640x480 or smaller and within the bitrate limit of the iPhone (over WiFi, at least).

I appreciate your response, but read the entire question next time you feel like putting on the grumpy old bloke hat :)

Thanks to those that took a moment for a constructive answer! Turns out that using the embed/object parameters slightly differently solves both the thumbnail and auto-orientation issues.

The bad thing is that the video object (at least using Quicktime - iPhone Safari doesn't support the HTML5 video tag yet) can't be scripted via Javascript.

The ADC has absolutely no mention of scripting the iPhone video object via Javascript - it doesn't say it can't be done, but the object doesn't fire any events that I can find either.

We use JS to handle tracking - we need to know how much of the video the user watched, if they muted it extensively, etc. due to the credit requirements of one of our clients. We'd love to deliver an iPhone-capable version of the site, but without scripting capability it won't happen.

So the bonus question to this post is does anyone have (or has anyone found) information regarding scripting the iPhone 2.x video object from a web page using Javascript?

That's awesome. If each of those custom tips cost $249.12 to make, then that cable is totally worth it.

As it happens, the video in this story is being served up by Limelight (a CDN similar to Akamai, but aimed at streaming media), not Netflix itself.

That said, if Limelight is pulling the original file to its edge servers (Los Angeles, in the case of this story) from a Netflix-controlled origin point instead of a Limelight origin point, then shitty speed on Netflix's server would translate to shitty speed from Limelight for the user.

Disclaimer: the company I work for (not Netflix!) uses both types of origin points with Limelight, and we do occasionally see serious speed issues on files served from our self-run origin points. Usually only on the really big ones though (like 1GB +). And a phone call to Limelight's client support usually fixes the problem within a few minutes.

Not really, no. Besides, only one row is on my land, the other row is on the neighbors land.

What I found most interesting in the discussion on Pidgin's site is that both the users and devs considered "arbitrary input resizing" to be a feature and not just a standard UI component. Personally, I consider being able to drag the border of that text input box a UI feature, not a program feature. Why would they override what most would consider the default (and expected) behaviour?