Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Do long pages screw up scrolling?

6 views
Skip to first unread message

Ted S.

unread,
Apr 20, 2013, 8:36:51 AM4/20/13
to
O12.14, Windows XP

<http://reason.com/blog/2013/04/19/boston-reopens-kind-of#comments>

Can't use the mouse wheel to scroll, or the keyboard (either the arrow
keys or the page up/down keys). Only the scrollbar on the right works.

And is there anything that can possibly be done to make reason.com not
be such a monstrous memory hog? 100% of the time that I visit the site,
Opera starts having severe memory-handling problems. It's reached the
point that I have to shut down Opera and restart it if I want to visit
reason.com, and even then there are still extreme memory-handling
problems. Right now I've got just two tabs open, and that's neough to
bring Opera to its knees. It's unbelievably irritating to have to shut
down a browser and reopen it just to visit a blog.

Oh, apparently what ever freeze I was having finished while I was typing
this message, because I can see dozens of mouse/keyboard movements that
were cached up during the freeze all be carried out over a few seconds.

--
Ted S.
fedya at hughes dot net
Now blogging at http://justacineast.blogspot.com

Warren Post

unread,
Apr 20, 2013, 10:08:38 AM4/20/13
to
I can't reproduce on Opera 12.15 x86_64, Linux. It took a while to load
(yes, it's a freaking big page). While loading, Opera had a lot of
latency and one CPU core was fully busy, but that's what I would expect.
Memory use isn't out of the ordinary. Nor do I see any significant
difference in how Opera 12.15 handles it vs. Firefox 17.0.5 x86_64.

--
Warren Post
http://my.opera.com/wpost/

JJ

unread,
Apr 20, 2013, 10:59:42 AM4/20/13
to
Rendering of large HTML file (100KB+) alone is one of Opera's weak points.
Even without styles/CSS and JavaScript enabled. Moving the cursor around
would require a lot of CPU processing. Unortunately, by design, it's not
multi-threaded, so the processing only use one CPU core even on multi-cored
CPUs. Enabling styles/CSS for large pages would noticeably decrease the
input response time and enabling JavaScript would decrease it even more
(especially if the page uses JavaScript heavily).

Unless you have fast computer (probably 2.5GHz+ with high performance RAM
modules), scrolling the wheel may not take effect immediately since Opera
may still be processing the change of cursor position for e.g.: which HTML
elements are under cursor (coordinate calculations), hover based status for
HTML, JavaScript events, etc. Check with Task Manager on CPU usage (and
split the CPU core view).

I'd suggest disabling styles/CSS and JavaScript for large pages. However, if
you intend to send data to the site like posting comments or logging into
the site, I'd suggest using other browser.

Robert Carnegie

unread,
Apr 21, 2013, 11:47:54 AM4/21/13
to
On Saturday, 20 April 2013 15:59:42 UTC+1, JJ wrote:
> Rendering of large HTML file (100KB+) alone is one of Opera's weak points.

I assume that you don't mean 100KB, because unfortunately that isn't
unusual for a web page that doesn't have a lot in it.
<http://www.bbc.co.uk/news/uk-scotland-north-east-orkney-shetland-22235784>
is 67 kilobytes to say that a cyclist was hit by a car (which I deplore;
the accident /and/ the web page size).

However, the computer may have frozen up due to some other cause -
particularly if there is only one processing core.

I find that the Google Groups interface to Usenet discussion can tie
up any browser in any Windows versionn - at least Opera, Firefox, and
Internet Explorer - when a discussion includes too many articles.
This seems to be a matter of Javascript running to construct a page,
since I get prompts from the browser inviting me to stop the script.
Google Groups recommends its own Chrome browser, but that doesn't
encourage me to use it.

What I haven't tried, and what also might be useful to you, is
Opera's "Turbo" feature, which filters the page content through
Opera's own servers before you see it and /might/ cut down on
the amount of stuff that's served to you, wanted and unwanted.

Or, people could avoid builting stupid over-large web pages.

Ted S.

unread,
Apr 21, 2013, 12:51:52 PM4/21/13
to
On Sun, 21 Apr 2013 08:47:54 -0700 (PDT), Robert Carnegie wrote:

> I find that the Google Groups interface to Usenet discussion can tie
> up any browser in any Windows versionn - at least Opera, Firefox, and
> Internet Explorer - when a discussion includes too many articles.
> This seems to be a matter of Javascript running to construct a page,
> since I get prompts from the browser inviting me to stop the script.
> Google Groups recommends its own Chrome browser, but that doesn't
> encourage me to use it.

Blogger does the same thing, which I find irritating. It also causes
memory handling problems on my system, although not as severe as reason.

> Or, people could avoid builting stupid over-large web pages.

No; web pages have to have all the latest anti-social networking crap.
That's something I find is another big problem. Reason specifically has
the problem of doing something to make Opera think all the page elements
are loaded -- I get the favicon and the dogear indicating that the page
is finished loading. But when I scroll the page to the point where the
Facebook/retweet/Google Minus buttons show up in the viewport, they
start reloading. I generally open pages directly to the comments,
especially if it's an article I've already commented on to see if
anybody's responded to any of my comments, and when I do the page search
on my username, that eventually brings up the buttons.

Thanks everybody for your comments. As it is I visit the site with only
cached images set to load, and plug-ins set to give that big "play"
button you have to click on rather than the Youtube-style preview image,
or the ads that actually play. Can't turn off Javascript since that
handles the comments.
0 new messages