Specifically, I'm looking forward to the WWW library being able to handle
multiple independent time/cpu-sliced HTTP operations. This way, a browser
doesn't have to halt all user interactions (save the HTTP interupt) until
after all HTTP operation is completed or interrupted.
With threading, for example, you would be able to read/scroll/following-links
in a page-screen, while another screen is busy doing HTTP "in the background".
Also, with multi-threaded HTTP, the <INCLUDE REFRESH=...> idea would be
more practical. Although I've implemented that dynamic feature in ViolaWWW,
at this point it's not terribly practical if the refresh rate is short,
because the UI will stop while the refresh is happening.
-Pei