And to partly answer your question: Oregon makes it a felony to sign someone else's name to the identifying envelope.
I think you'll be less enthusiastic once you understand how it works.
In my jurisdiction, ballots are stuffed into an image scanner (like a high speed fax) as they arrive, the votes are detected, any ambiguous votes are "electronically adjudicated" meaning workers alter the database to correct for voter intent or write ins. There's a nightly summary report, allowing people to peek at early results.
Add the problems with USPS losing 1% of all first class mail, and the lower end demographics being more mobile (changes of address), you get some real disenfranchisement issues.
It's true that vote by mail increases turnout, mostly with primary and special elections.
It's also true that vote by mail silently disenfranchises about the same number of people it enfranchises.
The correct solution is postal ballots for people who need them, poll sites for every one else. Thereby maximizing the number of people enfranchised and minimizing the number disenfranchised.
 In addition, we have locked dropboxes in many locations so you can avoid the USPS and the elections office entirely.
That's a numerical statement. Can you provide the numbers behind it? How much is turnout increased, and where did you get the 1% figure for the USPS losing first class mail from? I would assume that mail going to central locations are more likely to be delivered than to individuals. Same with issues related to change of address.
I filed FOIA requests with USPS, which they ignored. The metrics are done by a private third party, claiming the data is propriety (privatization allows govt to hide uncomfortable truths).
I got the numbers client lawsuit against USPS. Bulk mailers do their own metrics / tracking (using test mailings). They claimed USPS's "UAA rate" (undeliverable as addressed) was higher than claimed, so they shouldn't be charged as much.
Counties have the option to allow it, but most don't, so most individuals don't have this option.
Not that they seem to mind, though - I haven't heard of any complaints, anecdotally.
Is this different than how the count works in states whose ballots are delivered via volunteers with ballot boxes?
Most central counts have been using optical mark sense scanners, which are those multiple choice test reader thingies. Douglas Jones has posted online an excellent survey and explanation of various election equipment used. http://homepage.cs.uiowa.edu/~jones/voting/
My jurisdiction is very liberal about trying to count every vote marked, per voter intent. I understand that other jurisdictions will reject whole ballots if there's any problems scanning it.
With mark sense equipment, if I ballot doesn't read correctly, it's corrected. Ballots can be unreadable for all sorts of reasons, water damage, unfortunate paper fold, ballot printed askew, etc.
Full image scanner are newer. Votes are inferred using image processing (recognition), vs diodes firing off.
"Electronic adjudication" breaks the paper trail. To correct for voter intent, they're changing records in a database. Versus modifying / correcting ballots or ballot duplication (copying votes to a ballot which will then scan correctly).
covered by the Washington Post