A couple things should be noted here.
1) They're playing with fire. Big fire. You don't sue ~40 major companies and hope to get away unscathed; they'll make "InNova" pay in court costs, by dragging this out for as long as possible. That is, of course, if the patent holds at all in the first place.
2) The website. Anyone check it out? It looks like a 5 year old made it. For starters, it was made with a template, which means he probably made it himself. He didn't even change the comments/remove the from the source. Did I mention it looks like crap? Yeah, it does. Appears to have been made this year. Sketchy.
3) InNova LLC. The LLC is what has me curious here; besides the fact that this "company" has no products, its worth mentioning that he probably made this LLC simply for the lawsuit, so he wouldn't get burned if he lost. Characteristic patent troll.
4) More on "the company." If you look at the website, you'll notice they on'y have one product/solution. Can someone say patent troll? ONE "invention," listed as Patent #6,018,761. "This invention involves a system for adding to electronic mail messages information obtained from sources external to the electronic mail transport process. Applications include spam filtering for e-mail and news." If this doesn't point it out as what it is, then I don't know what does.
5) Why did he wait 15 years to file a case? Why didn't he file back when Gmail launched? If he really cared about the patent infringement aspect, then he would have filed immediately, circa his patent release, not 15 years down the road when anti-spam filters are very well established. He's clearly only in this for the money, not for his IP.
6) The patent itself. I don't like it. I'm not a patent lawyer though, but I personally find it a bit broad. He never explains how to pull off the context clues and analysis, or how/where to compare to lists/filters in a specific method, etc. It just seems really broad in its claims, which unfortunately is how these things roll. If he described a specific method, with code or something of the like, then I'd be more okay with the claims. However it just seems like he's generally covering the basics, without much detail at all. Anyone can claim that if you connect to ISP 22 and download messages to ISP 20, then compare to a list on Internet 18, that you can filter spam; the devils in the details.
Alas though, our patent system doesn't allow for such things, and while its generally pretty good in most areas, technology is where it really seems to stink. There's too many general, broad concepts, not enough specific details and methods.
EDIT: Also, yes, there's another problem. This "method" describes blacklisting and databases and context clues, etc... nowadays most spam filtering doesn't use the same methods contained within this patent... making it also questionable. There's no central database anymore that stores blacklisted context clues and whatnot, or email or organization addresses, at least in modern, self-learning spam filters.