[NgtCrwlr]

Well, my bright idea turned dim. I know I'm right track, but my Perl skills are just not up to this task. At least not without a hundred or so hours of head scratching and testing.

I thought it would be as easy as changing the decode to encode in this line:
decode_entities($f); decode_entities($s);

but for some reason it loops or repeats a bunch of times. You can see the output here:
http://www.mis-forums.com/vbbs/showt...threadid=17939

Then I started to look at newnews closer and realized that not only do incoming subject titles and poster names need to be encoded but they will need to be decoded before sending them back to USENET (I think).

Unfortunately, as we talked about earlier this is a problem of the way vB stores user names and thread/post titles/subjects. If vB stored the raw data (&,",<,>) and filtered it before displaying it like they do in all other data fields there would not be a problem.

This would go unnoticed for a long time because IE5 seems to digest the bad characters without problems, however, it is way bad HTML style to have &'s and greater/less than symbols where they don't belong, and the pages will never come close to validating.

Damn, I'm the bearer of bad news again!
I wish I could help more.