Re: lynx-dev LYNX: "go back to <prior page>" in OTHER browsers

[David Woolley]

> Now, all the time on various pages I see things
> (sometimes after every few paragraphs) like
> "go back to top" or some such.

I assume you mean previous page, as go back to top normally means
something else:  go to the fragment at the top of the current page.
Both have the same effect on the stack.

Particularly with previous page type ones, there are legitimate uses, and
I would go as far as to say that every page should at least contain a link
back to the home page.  The reason for this is that you can end up in mid site
as the result of a true web link (cross site) or a search engine hit, and
the only other way of getting to the root of the site is to edit the URL.

However, I suspect it is true that, whilst often achieving the above effect,
these are more often put in because the page designer has the pretension to
believe that he should override the built in mechanisms of the user's choice
of browser, and enforce his look and feel on all aspects of navigation of the
site.

An example I say recently of this was the use of something like:  
javascript:history.back() to simply simulate the browser's back button
(admittedly this was by the 14 year old son of one of the people I work
with and the page showed other signs of having gone mad with the design
(red on blue graphic buttons, dominating background, etc.)).  This doesn't
cause the stacking problem, but adds no value to the page.

One other legitimate use is as a browser bug workaround.  Some versions
of IE will show the correct destination in the tooltip for the back
button, but actually go somewhere else (forward on a parallel branch!).
We had to add a back button in the HTML to make force a safe destination -
I'm now getting pressure to put them everywhere, for a consistent style,
even though that sometimes means passing context to a CGI script that
it wouldn't otherwise need.

Conversely, one version of IE (3?) seems to break if you stack too many
forms.

Another reason that designers do it is in the forelorn hope of being able
to detect when you leave the site, so that they can free any resources
reserved for the session (both data resources and potentially reserved 
stock in a shopping list system).

> to unwind, causing and waiting for much internet traffic
> to regrab those pages that you only want to forget about
> anyway.

The only time you end up with internet traffic is when you pop an
absolutely uncacheable page (and note that Lynx caches rather more
aggressively than IE4 in this respect).  The increasing use of dynamic
forms pages does mean this can be a problem.