One observation that has surprised and concerned me as I watch the posts scrolling by is the substantial number of high reputation authors with very low follower counts... What are your thoughts about this?

I think this one is easy to explain - a lot of high reputation accounts have obtained their reputation by delegations or payment to a voting service. Therefore the (lack of) quality will contribute to their follower count. I was thinking along these lines as I read your criteria and wondered if you factor in bought votes into the algorithm somehow.

What are the important factors to include? Do you think I should drop any factors that I mentioned above or add other factors that I didn't mention?

Tough one - I've been thinking about this quite a lot too (in my efforts to filter out shit-spam-content) - In the realrobinhood posts, steemchiller references things like Resteems and comments. These are often good indicators that a post is "better" (although once again, the caveat is comments by spammers like krsuccess).

How can a method for suggested vote calculation be validated?

I missed something very important from you somewhere along the line so I don't know what the final objective of this program is?

As far as simply deciding a vote percentage goes, for me, this is an impossible question because I vote in clusters. Some days, I'll pop in a vote on 20 posts (I know, sub-optimal) and other days, I'll vote on nothing and leave my "back-up auto-votes" to take over. If it's the latter, then I don't mind whether it votes on 1 post at 100% or 10 at 10%. That might be just my mindset though.

moecki was correct to dynamically identify the index location (I do the same) - it's not bullet-proof as in one update, an index changed (I can't remember what it was, but it was along the lines of "comments" or "replies" becoming "children") so it's important to keep an eye on any planned updates either way.

I can't advise you about collaborative programming - any projects that I've worked on of this nature have had very clear delineation of roles (i.e. HTML / CSS / PHP, .NET, etc.). Obviously, the cleaner and more maintainable the code, the better but even with that, every coder will code differently so rewriting large chunks might be hard to justify. Perhaps something to consider if you think it will make your job easier, rather than worrying about what other people might do.