Okay, this time I just tested something on the command line, and it appears to work. At least for simple HTML with no confusing attributes in tags:

perl -le '$d = "<moose>elk</moose>"; $p = qr/<.*?>/; $d =~ s/$p/" " x length $1/ge; print $d'

The is even simpler than what I suggested previously:

• the regexp is very simple: make a non-greedy match from < to >
• no need for the '1 while' construct. Just use the /g modifier.

But once again, one of the HTML parsers may do a better job at this.