Yet another interesting Google/Mozilla connection not unlike the ones I pointed out not too long ago on my site. Great news, nonetheless.

Yeah, Google's gonna be the king of the Net, and its steel will be Firefox. We will take back the Web. :)

...and another good reason for me to avoid Google. My broadband connection has a usage cap, and I don't want to waste valuble bits by having sites I may or may not want to visit download everytime I do a search!

Now, I see that I could disable it by going into about:config, but why should I have to? Should it not be up to me to chose what gets downloaded to my computer without having to 'opt-out'? And could this not become a possible malware attack vector?

Dan: I presume you have images switched off too? I mean, why should you download bits you don't want.

Seriously though, people who are techy enough to be concerned about the number of bits they download are techy enough to switch the thing off.

There is more about this in this article on Slashdot.

- Save e a tree, eat more woodpeckers

It would be nice if Google can generate valid HTML... It is not THAT hard, is it?

minghong: I believe they strive for as few bytes of code as possible rather than standards compliance.

David,
I strongly disagree with your comment that "people techy enough to be concerned about the number of bits they download are techy enough to switch the thing off".
In the UK, a lot of people who are not heavy internet users and therefore almost by definition non-techy have capped broadband accounts. Typically these allow a monthly transfer limit of 1GB. I doubt that Google's current implementation would be enough to put such people over the limit however it is not a road I think we should be going down.
These is leaving aside my reservations about possible violations of acceptable use policies with regard to the internet in the workplace. I really don't see this as a good thing at all and without doubt should be an opt in rather than an opt out.

Does this prefetch feature really make browsing more efficient?

Using Firefox 1.0.2 and LiveHTTPHeaders I've watched as certain sites are prefetched. Do a google search for "stanford", or "microsoft" for example, and you can see the prefetch. But if you then click on the first link, it fetches the whole page again anyway. I would have expected the browser to use the cached version of the page, or at least put If-Modified-Since in the header.

Not that I'm complaining. I'm just curious how this is supposed to work?

Steve said: "I doubt that Google's current implementation would be enough to put such people over the limit"

Exactly. Why all the fuss.

I think I've figured out my own question. Stanford and Microsoft disable caching of their homepages, or at least fail to provide an expiration date.

If I'd tried my experiment with "mozilla" as the Google search test subject, I would have seen useful prefetching in action.

All the fuss is because it's sensible to discuss how these techniques might be used in the future and all assess their impact. Google's current implementation may not be significant in terms of bandwidth usage however future implementations by google or other sites that extend this may have a greater impact. As such I think it should not be a default option in Firefox and is worthy of discussion, just like we are having here!

I wonder why Google decided to use rather than . Prefetch is not standards compliant while Next is.

A clarification of my last post.
(link rel="prefetch" href="...") is proprietary
(link rel="next" href="...") is standards compliant

My company just started allowing Firefox -- but this prefetch thing will kill it because while we individuals may not look at everything in the cache, SOMEONE eventually will and we individuals will be held responsible for having chosen to go to whatever is in the cache.

HOW do I turn this OFF?
Please?

I don't want to tell management about this. I don't want necessarily to erase everything stored while browsing either!

Dan, this doesn't do it for *all* searches, just ones where it thinks you're actually likely to click on the first result.

Oliver, you're not quite right there. Read the spec. It was designed to be extended and there is nothing proprietary about our implementation of "prefetch".

- A

Yup, just like the meta element of HTML, there is no limitation on the value of its attributes.

"Dan, this doesn't do it for *all* searches, just ones where it thinks you're actually likely to click on the first result." - are you certain about that? That doesn't seem to be what Google is saying.

And I _do_ think this could 'tip the balance'. I myself am limited to 1Gb a month (but it only costs me something like $10/month, if my maths is right). Looking at Google sometimes the size of the first pages is quite large.

BTW I surf with all adverts and all Flash blocked (I love extensions). Pictures I compromise on - the background images on my favourite sites are blocked but most pics are allowed through.

Dan100 said: ""Dan, this doesn't do it for *all* searches, just ones where it thinks you're actually likely to click on the first result." - are you certain about that? That doesn't seem to be what Google is saying."

Um.. surely Google assumes you're going to click on the first result. That's why the result is displayed as the first.

If Google thinks you wouldn't want to click the first result, then why on earth would it be displayed at the most prominent position on the results page?

Furthermore, if Google would admit that their 'first result' is not likely to be clicked on, then they would be proclaiming that their whole "Im Feeling Lucky" option is useless.

To wit: we can safely assume this prefetching feature would work for *all* searches.

Just to clarify, that last post was in answer to Robert Morris ;-)