There are several problems with this page:

1) It sends a multipart/mixed content type that includes an incomplete MIME
message.  The message ends with the closing HTML tag, not a boundry tag as
it should.

2) It sends a bad charset:

Content-Type: multipart/mixed;boundary=xxyyzzzz0011294;
charset=ARRAY(0x94365ec)

3) The refresh URI isn't valid:

This is a multipart reponse in MIME format
--xxyyzzzz0011294
Content-type: text/html; name=Update
Refresh: 5,url=$back_url
Title: Updating

This happens to work in Netscape because it thinks that the URI is bad and
treats it as if the url= parameter wasn't there.

It's interesting to note that there are two problems with Mozilla at the
moment as well:

If you include anything after the content type in the Content-type:
paramter of a multipart/mixed doc it isn't parsed correctly:

Content-type: text/html; name=Update

I have a patch for this that will parse it properly but if you want older
versions to work you might want to remove the ; name=Update from the
content type header.

Also, I've discovered, much to my chagrin, that the Refresh: header doesn't
work at all with multipart/mixed documents.  It doesn't matter if you
include it in the http headers or the headers of the document.  This is a
less easy fix because it's an architecture problem but I'm working on it.

I have to ask.  If the document is being truncated after the first
document, why is it using a multipart document at all?