Linux Flash going to be limited to Chrome

Christopher Browne cbbrowne-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org
Sat Feb 25 16:42:45 UTC 2012


On Fri, Feb 24, 2012 at 4:37 PM, Lennart Sorensen
<lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org> wrote:
> On Fri, Feb 24, 2012 at 04:34:43PM -0500, Scott Allen wrote:
>> Ever consider creating a "to be read" bookmarks folder?
>
> I don't use bookmarks.  Ever.  Removing obsolete bookmarks is too much
> work, and tehre is no convinient one click 'add to bookmarks' either.

Firefox has a pretty sophisticated history browser, these days, and
the data is all captured in a SQLite database, so if you wanted to do
some of your own analysis, it wouldn't be terribly hard.

And I rather object to the "no convenient one-click add to bookmarks";
I have had that kind of functionality for "Post to Del.icio.us" for
*years* now.  All of the big name web browsers permit dragging buttons
into areas that allow 'convenient one click', so it *is* easy to get a
snippet of code into place to allow capturing and storing a bookmark.

Using del.icio.us has a few little issues:
- Requires a sign-in process, but that's *once* in a long while
- Stores bookmarks remotely
- Requires another interaction, as it pops open a window to allow
adding tags to the bookmark

I'd be shocked if it's not possible to create an analogous "click me!"
button that utilizes a local bookmark store.

I find local bookmark stores uninteresting for a different reason,
namely that I use several computers, and rather prefer the notion of
the store being shared.

Of late, Firefox does this "in droves" as its "Sync" notion allows
sharing history and bookmarks and the list of active windows across
multiple Firefox instances.  It claims to encrypt all the data before
uploading to the Firefox site used to control sharing so that they
don't know anything about my history; hopefully true.

> I expect my browser to not be wasting tons of rtam because a page is open.

At a fairly banal level, that may seem reasonable.

In principle, the "cost" of a browser page not being viewed might be
the bitmap image of the material.  Switch back, and the image can be
restored, and the image might even be reasonably pushed out as a
swap-like thing.

Unfortunately, that doesn't fit with the expectations of what should
happen when you switch *back* to the hidden page.

Expectations instead are that the state of the page should be kept.
- If I have filled in parts of a form, they should still be filled in
- If I have reshaped the window, text ought to reformat to fill the new width

In practice, today, a great deal of the web sites that people care
about (*.facebook.com, *.google.com, *.twitter.com) have increased the
footprints of their context to a spectacular degree.  Hugh observes on
the thread that he's got 58MB of "facebook.com" context in his browser
despite not having an account or any reason to be visiting that domain
directly.

It definitely doesn't cost "nothing," and the cost seems to be getting
increasingly spectacular.

This parallels Jim Mercer's comments at the January meeting on the
"bloat trouble with Java"; the trouble isn't with the language itself,
rather that unthinking programmers see that there's an option to set
the size of the VM to be "really big," and imagine that's a wise thing
to do.  Only to discover that since 7 other apps running on that
server *also* had programmers that had the same idea, the server is
now struggling because rather than them all fitting nicely into 256MB
of memory, they're swapping (unnecessarily!) across 8GB of memory.

The same thing has pretty evidently happened with the use of
JavaScript in web applications.  Each application behaves as if it was
the only one using resources, and so, shockingly quickly, some bits of
Facebook shared context (which is disturbing in some privacy-related
senses) get to add up to a shocking level.  An analogous answer to
what happens with Java might be appropriate, namely for there to be
restriction as to how much memory any given app gets to chew up.  But
that is mostly a "political" problem, and I'm not nearly sure how to
get there.
-- 
When confronted by a difficult problem, solve it by reducing it to the
question, "How would the Lone Ranger handle this?"
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists





More information about the Legacy mailing list