Why can't all ads be cool?

[Sirotnikov]

Uh, yeah?

The point is the browser rendering speed, not the speed of the internet connection (which is obviously variable)

Look, I think its a beautiful commercial with an incredible concept. Only its somewhat dishonest.

Things like cache, prefetching, connection handling, all affect real world performance, even in ADSL connections. Handling page reflows is just as important factor in precieved rendering speed. All of this is completely ignored, even though it is probably has a larger effect on user perceived responsiveness than net rendering speed.

But you know what? That would be fine if they hadn't pretended to enter the actual address (or at least google it) prior to launch in all 3 segments, thus leading people to believe they are watching real world performance. Which they are not.

Most commercials are honest enough to add "real world results may vary" when they show custom built lab tests.

A beautiful concept, but a misleading message.

P.S: Just watched it again. The mouse is not pressing "go", but pressing "forward", thus loading an already visited page. IIRC firefox holds several previously rendered pages in memory for quick rendering. Is chrome relying on the same trick for its speed demonstration?

[Asher]

Look, I think its a beautiful commercial with an incredible concept. Only its somewhat dishonest.

Only if you're a ****ing moron.

It's an ad about the speed of an internet browser, not the speed of your internet connection.

Reminds me of those misleading videos on TV of cars going really fast on private roads. In the real world there's other cars on the roads. MISLEADING!!!!

[Sirotnikov]

Only if you're a ****ing moron.
It's an ad about the speed of an internet browser, not the speed of your internet connection.

Is rendering of locally cached (or memory cached) copies a relevant portrayal of browser speed?

Reminds me of those misleading videos on TV of cars going really fast on private roads. In the real world there's other cars on the roads. MISLEADING!!!!

I think its more akin to a commercial demonstrating your car speeds in a world without air friction, pretending to do a cold-start but with the engine and tires actually warm.

Which would be misleading.

[Asher]

Is rendering of locally cached (or memory cached) copies a relevant portrayal of browser speed?

Yes. It removes external, irrelevant bottlenecks.

When you run browser benchmarks, they're all designed to fetch all necessary data before running anything. Why? Because any ****ing moron can understand that a user's independent network speed is a confounding variable. The browser cannot make web servers move faster, so why the **** would they count that?

I think its more akin to a commercial demonstrating your car speeds in a world without air friction, pretending to do a cold-start but with the engine and tires actually warm.

Which would be misleading.

It would be, but it's not at all like that. The browser runs like that in the real world. Hell, the technology behind Chrome is what makes WebOS work (all read locally) and how ChromeOS will also work (all read locally).

The Chrome browser itself can also be used in the intranet or local environments.

Internet latency is nothing Chrome can improve on (well, it tries to with stuff like DNS pre-fetching but this is something not relevant), so it's completely ****ing irrelevant when comparing browser speeds. What's important in the browser is how fast it renders, since the internet connection is the same ****ing bottleneck in every single browser. It's an external factor unrelated to how well the browser performs.

Not to mention the incredible ****ing variance in internet speed. If Google ran this test on an OC-192 connection 100 feet from the server of the website would it really make you feel better?

Stop wasting my time with this ****.