More likely someone's hiring them to spam every SMF instance they can find--the Compendium doesn't have high enough traffic that I can see anyone bothering to attack it specifically.
There are a lot of other things that could be done about the spam flood, with varying levels of effectiveness:
-send new users' first posts to a moderation queue for approval, if they contain more than x links or consist only of a link (4-5 is probably a sane value for x. Requires moderators to catch the occasional legitimate post.)
-incorporate an obfuscated timestamp field into the registration and posting forms, and don't allow the submission through if the timestamp is more than several hours old. This defeats one common type of bot that just plays back a previously crafted HTTP POST message.
-incorporate a field in the registration or posting forms which is rendered invisible by CSS and refuse to process the form if it is filled out. Bonus if the name of the widget inside the HTML makes it look like it really should be filled out. This defeats another common type of bot that just fills out all the fields in the form with something that looks vaguely plausible. (You also need to include a hidden instruction for people using screen readers so that they understand what needs to be done.)
-kitten captcha: a type of captcha where you have a randomly ordered set of 8-9 images that look similar to a computer but are easily distinguishable by a human (the standard set is small fuzzy animals--kitten, puppy, bunny, etc.), and ask the user to click on a specific one. Less painful for the average user than a standard distorted-letter captcha, and most bots aren't pre-programmed to deal with it. (Doesn't play well with assistive equipment for the blind, though.)
