And here is the answer:
<noscript>
<META HTTP-EQUIV="Refresh" CONTENT="0;
URL=Callback?JAVASCRIPTSTATUS=NONE">
</noscript>
<snip>
This testing strategy does not answer your stated question. It may give
an indication that a client has scripting disabled/unavailable, but only
when that client also supports, and acts upon, <META
HTTP-EQUIV="Refresh" ... . As many modern browsers allow META refresh to
be disabled independently of (but obviously in addition to) javascript
the strategy risks giving an indication that scripting is enabled on the
client when in fact it is not. Generally erring in that direction should
be less desirable than erroneously assuming scripting is disabled when
in fact it is available.
However, the formal rules of HTML categorise META elements as
"head.misc" and makes the only valid location for such an element a
direct child of the HEAD element, while NOSCRIPT elements are "block"
elements and may only be descendants of the BODY element. Thus in valid
HTML a META element cannot be a child of a NOSCRIPT element. The two are
only allowed in mutually exclusive contexts. As a result this code will
also potentially give a false indication of client-side script support
on browsers that take a more literal interpretation of HTML, or apply
realistic error-correction rules.
It is always a bit depressing to be reminded of the actual number of
individuals working in the production of HTML who have little or no
technical understanding of HTML.
With the NOSCRIPT element in its correct context within the BODY it
should still be possible to make a GET request to a servlet. Placing an
IMG element, or an IFRAME element within the NOSCRIPT element would
result in valid HTML and encourage the browser to make the desired GET
request. But again, false indications of client-side script support
would be generated by browsers with images turned off, or incapable of
supporting IFRAMEs.
My sincere thanks to everyone who looked at my inquiry,
even though nobody responded with a constructive answer.
According to my newsreader you have allowed less than 13 hours for a
reply. On an international group it takes 24 hours for all of the
contributors to even see the original massage, let alone respond. But it
is quite likely that you haven't received any "constructive" responses
before now because everyone knows that there is no 100% reliable method
of acquiring the information that you have stated that you want, and the
chances are extremely slim that the information you claim to want is
actually the information you need. Though with little browser scripting
experience it is unlikely that you will see that, even if someone took
the time to explain why.
Richard.