W3C home > Mailing lists > Public > public-web-perf@w3.org > June 2011

Re: Further concerns with PageVisibility

From: Kyle Simpson <getify@gmail.com>
Date: Wed, 15 Jun 2011 13:15:57 -0500
Message-ID: <C82DCC50A7794B98A693083729C6753A@spartacus>
To: "Sreeram Ramachandran" <sreeram@google.com>
Cc: <public-web-perf@w3.org>
>> If I have some stuff I want to show to users, and *they* say they don't 
>> care
>> (by opening my page in the background tab), I'm ok with that. If the 
>> browser
>> gets to decide that my stuff is unimportant to show to the user, I'm very
>> not-OK with that. Which means I have to now care about delaying such code
>> for PageVisibility.
>
> The browser is the user's agent. If you can tolerate the user choosing
> to ignore your page's initial animation, then you ought to tolerate
> the browser doing so, too.

Umm, you're not serious are you? There's tons of precedent for things that 
we trust only users to do/decide, and not browsers. If we implicitly trust 
the UserAgent in all cases, then why on earth do we have preferences which 
can affect the browser's behavior (such as caching, cache-size, transmitted 
UserAgent string, etc)?

The user and the browser are decidedly not equivalent (even when it comes 
just to browsing). I'm asserting my subjective opinion that in this case, 
one's decision is ok and the other's is not. But it's absurd to suggest that 
there's no difference between the two at all.


>> Moreover, if pages start doing that (delaying stuff for PV) widespread, 
>> it
>> will actually eliminate a user's implicit choice (which they currently 
>> have)
>> to have such "this site is loading" stuff (which they don't care about)
>> happen while the tab is in the background.
>
> Not true. They can continue to middle-click the link (to an
> already-prerendered page).

I think you either mis-understand the feature, or my assertion. My point is, 
if a user wants to ignore such "loading" stuff, right now, they can open a 
tab in the background (middle-click or whatever), and just wait.... And when 
they finally do activate the tab, the page will be sitting idly by (but 
already fully rendered and all the "loading" behaviors long since having 
run).

BUT if that site starts delaying such behavior until PageVisible, that event 
will not fire until that tab is actually activated/shown, so the user will 
switch to the tab a couple of minutes later, and boom! they'll see the 
behavior they were hoping had just silently passed by in the background 
already.

I don't forsee any way that a user could force a page to open in such a way 
that PageVisible fired, but that they in fact weren't bothered by seeing it, 
except perhaps if they map the middle-click to "open on my second monitor 
which I have turned off" or something like that.

Now, is there strong evidence that users are bothered by such loading 
animations/behaviors, and so THAT's why they intentionally use background 
tab loading to avoid them? No, I doubt that, obviously. I think the much 
more common case is they simply don't want to "waste" the 2-8 seconds most 
sites load/render, so they let that stuff happen in the background tab.

But, the point remains, right now they have a choice, and if a site 
implements PageVisibility delaying, they will lose that choice.


--Kyle

 
Received on Wednesday, 15 June 2011 18:17:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 15 June 2011 18:17:32 GMT