If you need to know, you either have to use some heuristics (for instance, it "validates", either against a DTD or the more usual "my browser doesn't barf on it"), or ask the user.

Impossible. At best, you can take a guess. But you can guess very reliably because HTML must have an HTML element.

If you don't know if it's text or HTML, then you're surely dealing with bytes, so you need to handle UTF-16le, UTF-16be, UCS-2le, UCS-2be, UCS-4le, UCS-4be:

/<HTML|<\0H\0T\0M\0L|<\0\0\0H\0\0\0T\0\0\0M\0\0\0L/
[download]

If you're somehow dealing with decoded text:

/<HTML/
[download]

Update: No, that's still not good enough. A text version of this very post would fail, for example.

But you can guess very reliably because HTML must have an HTML element
I forgot to mention that the html entered may be just a part of HTML, so assuming presence of "<html" tag will not work.

This is HTML:

Please use <code>...</code> tags around your code.
[download]

This is text:

Please use <code>use strict;</code> in your code.
[download]

How can one possibly correctly identify them programatically?

PS - This is the reason Atom is better than RSS. RSS doesn't provide a mean of specifying the content type, so it can't distinguish between text and HTML content. Clients have to guess. You could take a peek at how RSS clients do it, but I suspect they might work with less ambiguous content than you.

#!/usr/bin/perl -l

use strict;
use warnings;
use Text::FromAny;

my $log = '/root/Desktop/text.log';

open STDOUT, '>', $log;
my $entries= "<TITLE>Page 7</TITLE>";
print $entries;
my $tFromAny = Text::FromAny->new(file => $log);
print $tFromAny->detectedType;
close STDOUT;
[download]