[Tfug] Browser based UI's

Bexley Hall bexley401 at yahoo.com
Tue Jul 14 10:34:26 MST 2009


Hi,

I've been trying to give browser based user interfaces
"the benefit of the doubt" as potential general purpose
user interfaces.  But, I think I have come to the
realization that they just don't (can't?) work for
*all* types of applications.  I.e., Google's idea
that they can replace the desktop with web based
applications is  just a wet dream, IMO.

But, I have been unable to codify *why* these "just don't
work" -- whether it is the browser's "fault" (i.e., capabilties
that are missing in that type of interface) *or* that of
the applications, themselves (i.e., could these work if the
model adopted by the application were different than it
*has* been?)

Of course, the responsiveness of the network plays a role.
But, pretend you were connecting to "localhost" for all of
your "web based services"... so transport delays, dropped
packets, etc. were not an issue.  I'm sure you could still
see how many "applications" just *suck* using this form of
interface.

[i.e., imagine rewriting KDE so you opened a giant browser
window in your startup script that was pointed at http://localhost/kde
to provide that "environment".  Some of it would work "OK".
Some parts would be "passable, but not ideal".  And other
things would just suck horribly!]

Any thoughts as to what aspects of the applications/browser
make for *good* interfaces vs. bad?  I.e., obviously a 100%
static application can be handled reasonably well in a browser...


      




More information about the tfug mailing list