First, Mr. Clark has been doing a good job of answering your questions. I only comment because I have a perspective that I don't see here yet.
Second, Lemelson is certainly a larger-than-life figure in patent law -- perhaps the single biggest name in all of patent law. Some people hail him as a god; other curse his as an anti-christ. I suspect that the law was never intended to produce the consequences he wrought.
I think your probing of the enablement issue evidences a concern that perhaps some patents out there shouldn't be -- particularly in the software arena. I suspect your concern about enablement of software patents is a bit misplaced. Not that all software patents are properly enabled, it's just that I don't see enablement as a particular weakness of software patents.
Remember, you only have to describe the invention well enough that someone of ordinary skill in the art can make and use the invention. In some cases, a single sentence can accomplish that. The fact of the matter is that software is so malleable that it can be configured to do just about anything. In some circumstances, merely mentioning the intended result is enough to enable somebody to make and use something in software.
Here's an example. I'm hoping to build a fee estimate calculator on our web site for applications claiming foreign priority on another patent (that job is sufficiently clerical to give fairly good cost estimates). I've done some CGI programming (built a weight and balance calculator for private pilots using CGI/perl). Is there any doubt I can do it (or at least that one of ordinary skill in the art of web-based programming can do it) without having actually accomplished it? I don't think so.
So, in short, I think most software patents filed today are actually well enabled -- with or without a working implementation.
However, that malleability has an opposing edge -- it can undercut the non-obviousness of a particular invention. In fact, the Patent Office used to take the position that all software was obvious -- just computer instructions stored on a disk and that's been done many times over. Of course, that view didn't survive in the courts.