hellonearthisman: I mentioned this on http://computerworld.co.nz/news.nsf/fryup/fryup-three-of-the-best about how the Orcon cache is breaking youtube. And they are still doing it.
If I watch a YouTube clip in my browser and then open a new tab, paste in the same URL, I get sent another copy of the video when it should use the copy I have in my browsers cache.
I hate you.
muppet: No matter what I tried though, squid won't cache them. A bit more research shows that YouTube do some tricks to stop this working!
YouTube specifically implement several 'features' that prevent their flash videos being effectivly distributed by caches. There is even a (was GPL, now commercial) squid addon that enables caching.
freitasm: Correct. That's why ISPs can't just cache YouTube, instead having to "license" a Google box for this purpose - exactly the exercise TelstraClear has just gone through.
I hate you.
muppet: I don't quite understand this.
No matter what I tried though, squid won't cache them. A bit more research shows that YouTube do some tricks to stop this working!
insane: I'm surprised no one has asked the question why someone would watch a youtube video over and over again, each time closing their browser and re-visit the same video.
You also have to remember that the first time you viewed the video it would have been accelerated as a result of your ISPs cache. I'd rather pay for the traffic of the videos I re-watch later twice than having every youtube video I watch slow the first time.
I'm still curious about the correlation of an ISP having a youtube cache means that browser caches don't work!
insane: I just tested this (my ISP makes use of the Orcon google cache) and I closed and re-opened the youtube tab a few times, even closed firefox and re-opened and re-loaded the tab and the video was not downloaded a second time so there is no problem at all.
The only time the video will re-download is if I skip forward in the video, then it will pull down from that point on again.
Browser cache size set to 300MB
hellonearthisman: Back in Sunday, June 14, 2009, google released this message.
hellonearthisman: Thanks insane for your testing and others for your feedback.
Insane, if the Clip is fully buffered then there adjusting the timeline position shouldn't cause a reload.