- Build your entire page, up to but not including the terminating </body> tag. [*]
- You need to be able to flush output from the server. Browsers may not update the page if they do not receive a full line, so be sure to include a terminating newline in your output.
See Also edit
Examples currently work with: IE 5.5 and 6, Netscape 6, Mozilla 0.9.5, ...Examples do not work with: Netscape 4, Opera 6, ...
Arjen Markus: I have been experimenting with something akin to this - using the browser as a device for "on-line visualisation". That is, a long-running computation produces an output file, a plotting program produces a GIF-file from the last record in the output file and the browser regularly checks if there is a new picture. It is not fool-proof as yet, but the principle seems to work.CL responds that, yes, there are both browser-pull and server-push applications of the technique. Its kernel, to my mind, has to do with DOM navigation or computation. A few more comments appear in .[email@example.com]: This is a cool idea. I have been working with a similar concept to update a controller page with small agent.php windows that send data back to a controller page while running in the background. I made a mini test using straight html: http://www.nectechmedia.com/auct.sample/controller.htmstill buggy but, works- Matt
<meta http-equiv="refresh" content="10">With that you have to re-create the whole page, though.