More comprehensive reform would probably include a ban on patents in software, or at least a much, much shorter patent period for software and similar domains of invention that move much more quickly than, say, pharmaceutical development and approval.
For a patent covering a trivial software idea, it's ludicrous.
I'd be happy with a system that gives five years by default, using the current process. Twenty years only if you can easily prove non-obviousness and significant research investment.
Yes, the SHIELD act is treating the symptom, not the cause.
Such objective evidence could consist of successful publication in a peer-reviewed journal, for example, or of providing published references that say that the problem being attacked is known to be difficult, or of showing that others have tried to solve the problem and failed.
In short, the burden of proof needs to be on the applicant to prove nonobviousness, not on the examiner (or the defendants in a lawsuit) to prove obviousness.
If the USPTO is willing to grant ridiculous patents that it shouldn't, then the SHIELD act would seem to not make much of a difference at all. It seems to me that the prior art should be a non-issue to produce considering people have been making tutorials on how to make podcasts since '04/'05, but as long as the patents are being erroneously assigned, the SHIELD act will provide very little protection from these types of patent trolls.