I found a way to shrink the monster class by 500 lines or so, which isn't bad. While implementing the rest of the features, I made the following observations:

Observation 1: Some methods are highly specialized and have unidirectional dependencies. For example, the methods for inserting, updating and deleting blog entries require a slew of other methods, but aren't required by any other methods in the class. They essentially act as interfaces to the outside world (i.e. they're what would be declared as public methods in a real OOP language).
Observation 2: Such specialized methods are isolated. For example, the methods responsible for adding/updating/deleting post entries are totally separated from, for example, the methods that add/update/delete links. A PHP page or RPC method that does stuff with post entries will most likely not do anything with links, section, etc.

I was then able to isolate 5 groups of methods that fit the above observations and create child classes to the original (now shrunken) monster class to hold the specialized methods. When instantiating an object, the calling script now has a choice of 6 classes, depending on which functionality is required.

As reported on Slashdot, SCO has apparently shown analysts evidence to their claim... a whole wopping 80 lines of code. But that got me wondering. Even if the code was stolen from UNIX, and even if we assume that SCO owns copyrights to the code in the first place (a big if from what I've seen), is that really worth billions of dollars? Isn't that kind of like Wal-Mart suing some shop lifter for damages equivalent to years' worth of revenue?

Assuming that SCO is right, I can understand how they should be compensated for what was stolen from them. However, the compensation should be proportional to the actual damages. For example, if the estimated revenue from UNIX over the last, say, 10 years would've been $2.5 billion, they should receive $2.5b x S/T (where S=number of lines stolen, T=total number of lines in UNIX). Alternatively, they could be awarded a fraction (calculated the same way) of revenue from copies of Linux sold. Either way, my guess is that since UNIX and Linux contain lots and lots of code (what exactly falls under "UNIX" or "Linux" code anyway? The kernel?), what they deserve would be significantly less than what they're asking for.

It also ocurred to me (I'm sure I'm not the first) that one possible reason SCO doesn't want to disclose evidence to their claims to the public is to prevent Linux developers from coming out with a clean version of Linux. Considering how the code in question doesn't appear very substancial, I would be surprised if it took the Linux community more than a few days to swap all stolen code with new original code. Companies that use the "contaminated" versions can then upgrade to the new clean Linux, and avoid paying SCO any royalties. Well, maybe they'd have to pay for having used it in the past, but that's kind of difficult for anyone to prove, isn't it?

Anyway, all of this still depends on SCO's claim being true and them being able to prove ownership of the code in question. Those are still fairly large ifs and hopefully it won't come to much nastiness.