[extropy-chat] Firefox 2 [was: The End of Science]
robert.bradbury at gmail.com
Sat Oct 28 01:01:57 UTC 2006
On 10/27/06, BillK <pharos at gmail.com> wrote:
I'll make this brief so hopefully people will see I'm not trying to cause
trouble (though I realize it may be hard to know when I am... :-))
> How can you possibly remember what you've got open in 100 windows / 700
I have a big mind? (partially true). Though right now I'm using a 22
window, 125 tab session because the larger session was so slow it was
unusable (strange how 100 *minimized* windows require anywhere from 40-90%
of the CPU time -- one would think that good programmers would understand
things like "queues" or "interrupt driven functions").
My work strategy is to go to page summaries (/., Physorg, Science Daily,
PNAS contents emails, etc.) look through them to find what is interesting
and open those pages in tabs. Downloading a page can take seconds to
minutes depending upon the DNS lookups, network bottlenecks, CSS pages,
images, etc.) . So I open a lot of them quickly and return to reading
them when downloading is complete (this is a simple time optimization
strategy) and something else is going to take a while (linux builds [20
minutes], firefox builds [an hour?], etc.).
It is not unusual for me to be working on multiple areas of interest and to
have windows and tabs open dedicated to each area. I generally separate
them into workspaces under Gnome (I have a personal (email/social network)
workspace, a "corporate" workspace, a "news of interest" workspace, a
medical research workspace, a Linux development workspace, a Firefox sucks
and someday I'm going to determine precisely why workspace, etc.). Because
the machine functions as my web & email server I do not like to reboot it
unless absolutely necessary. Because Firefox takes so long to restore its
former session state (10-20 minutes with a lot of windows and tabs) I don't
like to close and restart it either. If I open 15 windows and 50 tabs a day
after a week I'm pushing the numbers mentioned above.
If Firefox had better page caching strategies (Netscape 4.7 allowed control
of this) reopening pages could be nearly instantaneous and I might be more
inclined to limit my session state size. But it doesn't. If it had better
heap management it wouldn't be such a memory hog that it requires people to
increase the memory size on their machines (its *supposed* to be *free*) --
to date its cost me $80 and it looks like I may need to buy another GB.
And *NO* program which is in "production" should crash the way Firefox does
. Do you want the software handling the data being read out of your
frozen brain being slowly disassembled to crash during the middle of your
upload? I think not.
Until the problems I've mentioned are fixed its a *toy* and not a *tool*.
1. I am not working in a sub-slowsky environment -- I've got a DSL
connection and the machine is a Pentium 4 with 1.5 GB of memory (after
upgrading it from 512 MB which Firefox ran out of quite quickly).
2. They are *not* catching memory allocation failures in C++. A robust
system or program will tolerate hardware failures (say a swap drive dies
and/or no more paging space is available). This is not the case with
Firefox as it is currently written.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the extropy-chat