Browser Histroy

[prev] [thread] [next] [lurker] [Date index for 2003/08/13]

From: Mark Fowler
Subject: Browser Histroy
Date: 10:47 on 13 Aug 2003
This is my rant.  There are many like it, but this one is mine.

Browser histories don't work.  They come from the day where someone would
open _one_ window and browse with that _one_ window.  Thus they all store
the time the person loaded the page as this translates to the order that
the user was viewing the page.  So when someone wants to reopen that page
that they were at ten pages ago they can do - they just look back though
their history and it's the $history[-10] entry.

Fast forward to a few years later where most browsers have multiple tabs
(and even the crippled ones have multiple windows) and suddenly this is no
longer true;  The order in which pages are loaded no longer represents the
order in which pages are viewed.

To illustrate this, imagine a site like http://www.ntk.net/.  It's a site
with links on it to many other sites.  Typically, I view such a site by
opening each link on it in a new tab, and then exploring the other site.
When I'm done I close the tab and look at NTK again.  Now what's the last
thing I looked at?  The site that was linked off of NTK?  Nope.  It's NTK
itself.  But I loaded that _ages_ ago, so it's way back when in my history
file, buried under the other hundred or so pages of the linked sites.
If I close this window and I can't remember NTK's url, what's the chance
of me being able to find the site that I looked at again?

Another situation - one that I come across quite often - is that while I'm
working I'll google for some documentation that ends up sitting on the
virtual desktop next to my text editor for many many days.  However,
during this time period I'm highly likely to open up another browser
window when I'm due a break and check the news, check out things people
have been posting to irc, read some cartoons - anything to clear my head
for a minute.  Despite all this browsing at some point I'll complete the
work and close the editor and the documentation I searched for.  The
problem comes when I come back from my tea break and someone tells me that
there was a bug in my code;  Suddenly I need to get back at that
documentation that I thought I was done with (so I didn't bookmark it.)
No-one's touched the computer since I closed the window - it was the last
page I was looking at, but I'll be damned if I can find it in my history.
It's buried under the hundred or so sites I've looked at in the two days
since I opened it.

So what can we do about this?  Well, it seems to me it would be nice to
record when someone was _done_ with a page rather (or as well as) when
they started looking at the page.  The sensible way to do this would to be
to simply record when a page was closed as well as when it was opened.
It's not that hard (assuming you have proper session recovery when your
browser unexpectedly exits like Galeon et al do.)  So why don't browsers
do it?

This is my rant.  One of these days I'll STFU and patch a browser (or
someone will tell me there's a browser that does this.)  Till then, I'll
continue to spread the word.

-- 
#!/usr/bin/perl -T
use strict;
use warnings;
print q{Mark Fowler, mark@xxxxxxxxxxxxxx.xxx, http://twoshortplanks.com/};

Generated at 14:02 on 01 Jul 2004 by mariachi 0.52