These days with the way hardware and software are intertwined, you never know. 30 years ago I remember a geek practical joke running around about software being able to fry your CPU, and it was a practical joke because those who really knew the system knew you couldn't do that (crystal-based timing). Nowdays software can easily overclock your CPU or GPU and fry it.
Obviously, software can't fix an issue like the EVO 4G coming apart at the seams. That's pure hardware. But if the iPhone's antenna is software-controlaled to a large extent, it can be fixed. Apple was bragging about its new method of getting better reception, intelligently switching around and such, and that sounded entirely software. It also sounded very complex, the perfect place for a bug that causes the problem.
I agree.
I base my skepticism on the description of how the problem is generated by hand placement. Hand placement isn't software, it's hardware. It's touching the antenna, which any engineer will agree will change the level of the signal, whether receiving or transmitting, by de-tuning the resonant characteristics of the antenna at the frequencies of interest. If the problem is bridging between the two antennas, it gets much more complicated, I imagine.
The actual nature of the problem is that hand placement causes the signal level to drop. Period. That's not software, it's physics.
Now, granted that there are surely some fancy algorithms in the software which deal with widely varying signal strengths. There have to be. And probably there are some fancy adaptive functions that deal with all manner of issues from orientation to ghost-suppression to echo-cancellation and so forth, most of which are way over my head.
But ultimately, the cause of the problem is hardware -- a hand on an antenna. At best, the software will attempt to accommodate (live with), or compensate for, the hardware problem. It cannot correct it, because at the root, it is a hardware issue.
That said, I know firsthand of a large number of hardware issues that were accommodated or compensated for in software, where the symptoms disappeared completely. It all depends on the hardware problem. Some, you can do that. Others, not.
Conference in half an hour.
That has been my opinion since a Freeper commented that from his house his iPhone occasionally grabbed the weaker signal from the two nearest towers. When you are comparing signal strength from a number of competing sources, especially when you are on the move, and then choosing when to lock in to which, change to another while dropping the first, while all the variables are constantly changing, seems like a pretty tricky proposition with lots of opportunities for dropping a signal. If my description is even close to accurate, I don't see how it works at all.