> You're trying to make the browsers do something they weren't designed to...

I most certainly am not. Browsers are "designed" to implement the HTML, HTTP, and Javascript standards. Others too, but these are the three I'm using in RinkChat. No part of RinkChat relies on exploiting functionality not covered by these standards, and therefore it should work without flaw.

> Does RinkChat fit in with the HTML 4.0 spec?

Yes, but only the <br clear="yes"> tag, which IE 3 does not implement properly, is something I particularly depend on, as the rest of the HTML I use is pretty generic. The HTTP/1.0 and HTTP/1.1 standards are more important, as these are what indicate, for example, that the server, not the client, should be the one to close an HTTP connection; that the HTTP Refresh header should be given in a certain syntax and cause certain behavior; and so forth.

The Javascript commands I'm using are all standard ones that the browsers I'm trying to support claim they do. IE 3 doesn't support any, so I don't bother with it. Netscape 3 supports all but one, so I special case that and don't send it to Netscape 3. (Which is a type of special case that doesn't bother me -- backwards compatibility issues are to be expected. Anyway, the great thing is that Netscape 3 actually works, so it doesn't even NEED that particulr command, which is only used to work around bugs in other browsers.)

So here's the thing. When a browser receives a <br clear="all"> tag, it is supposed to render everything up to that point and display it on the screen. When IE 5, 5.5, and 6 fail to draw the text properly and you have to highlight it with the mouse to get it to show up, that's a bug. When you do a manual reload in Netscape 4.x and it doesn't render when it hits the <br clear="all"> tag, that's a bug. It's not something that browsers are not "designed" to do. What they are designed to do is to render the damn text on the screen.

When a browser that claims to support the window.scrollTo javascript method, what it is designed to do is to support it. And so when a browser is told to scrollTo the end of the page, and it only scrolls down half way, and the browser requires scrollTo and scrollBy commands, dispensed in 0.3 second intervals three times before the browser acquiesces, this is a bug. It's not something that it isn't designed to do.

It's not a great shame and dishonor to have a bug or two in a software application. But browsers are categorically laden with them, and it was supremely disheartening to test a frankly simple use of widespread standards and not find even ONE browser, out of a dozen or so, that worked as it should, and it was worse to find so many that didn't work at all.