> ..., DW offers "Western European" as an option, and if I
> choose that then ISO-8859-1 is what actually gets inserted into the
> (mostly useless) meta-tag. So that makes 3 synonyms for the same charset.
>
> Also, after thinking a bit, the fact that the output originates
> dynamically as php shouldn't matter, right? If I were to download my php
> page, save the source as *.html, and then publish that file as *.html,
> then the end result is the same - as far as DOCTYPE goes.

Not entirely. The best way to send the character set is _outside_ the
document itself. It is a strange that you should parse the character set
used out of the encoded document. What about utf-16? or utf-32? Could
you parse those character sets as easily out of a documented that is
written in it? Off course not. And if you can, there is no reason to do
so anymore.

So character sets are sent in a header. PHP supports this in two ways:
- set the default document type and character set in your PHP.ini,
- send a header your self with the header() function;

Especially the first way can be tricky is you do not realize it. Your
server may be sending "Content-type: text/html; charset=utf-8" in a
header, while your code may contain "Content-type: text/html;
charset=iso-8859-1" in a meta tag. Which of them is true?

> So this was really an HTML question, independent of whether that the
> output originated as php.

As you saw above, not entirely. Plain static HTML files have no control
over their headers.

Best regards.