We decided to run a witness because we wanted to support the network and we think we're highly qualified as a witness, especially when there are problems.

We wrote one of the key pieces of the Steem code: the peer-to-peer network layer that enables steemd nodes to communicate with each other (i.e. share transactions and blocks). This code was originally written for the graphene blockchain library and re-used in the Steem codebase and other graphene-based chains, so it's actually used in a number of active blockchains nowadays. We've also developed other code for steem under various contracts.

It's actually pretty challenging for witnesses to detect most code bugs in the Steem code ahead of time. To do a good job, the testers should really be interacting daily or at least weekly with the programmers doing the code as they need to understand the nitty gritty details in order to make good tests. This really doesn't fit with Steemit's current coding model, so I really feel the onus is on Steemit to up their testing game. About the best I think we could do as witnesses is ask to see what tests they've created and the results, then try to make a judgement about if the testing is adequate. Of course, we can also look at the code, etc, but it's unlikely that such an examination will detect anything Steemit hasn't already checked.

We promote it a lot on our web site: https://blocktrades.us. And indirectly through any marketing that leads people to our site. We also actively upvote/fund reasonable proposals that people make to do external promotion of the Steem blockchain.

We have and do support various communities via targeted upvoting, especially when they are just getting started. We have also occasionally delegated steem power to such communities.

As long term graphene coders, we're about as qualified as anyone is to read the Steem codebase. We have a large team of experienced C++ coders (our most experienced guys have 20 -30 years experience in C++) who've worked on extremely challening code.

We've developed various Steem-based applications, but we haven't released them for public use yet. The ones that come to mind off hand are a messaging system that allows chat/email over the Steem blockchain and an accounting system that can analyze Steem wallet transactions for tax purposes (but this is part of a much larger system). We also have a prototype vote bot that we've never really used.

Communication from Steemit is a real problem, but I don't real feel the same thing is true of the witnesses. The solution to this problem, IMO, is for Steemit to post more often.
So I guess I don't think it's a great idea, as I think it's just likely to stifle honest dialog in the channel if witnesses feel like they have to monitor every word they say.
The discussions aren't really that fascinating most of the the time and generally it seems like anytime there's an idea that gets any traction in the group, one witness or another winds up making a post about it to get the opinions of the wider Steem world. Some witnesses I would say are pretty much elected based on their work publishing such ideas on the blockchain (which is very reasonable to me).
That said, I'd be fine if the discussion itself was made readable, it's not a big deal to me one way or the other, and maybe it would end the conspiracy theories that seem to arise around the channel (or probably not, since I think there would just be the claim made next that there's another more secret channel, lol). So probably it's a pointless exercise.
Oh, one other issue is that sometimes there are real security issues discussed in that channel (i.e. a zero-day bug that's not yet patched). There really needs to be a secure chat channel for such communication to prevent hackers from taking advantage of that kind of information before witnesses can coordinate to update their code. This is obviously a rare situation, but it does happen.

Steemit has the funding and most of the experienced steem developers (well, we're probably the one exception to that rule).
Witness pay was cut along time ago to the point where it's really not practical for witnesses to pay for programming teams. Even when our team was doing work on the Steem blockchain, it was under contract with Steemit itself, as our witness pay simply couldn't cover the programmer's salaries.

We have the personnel and experience to do the main tasks that I think witnesses can reasonably do: 1) produce blocks, 2) respond to emergencies, and 3) evaluate proposals for new features in the blockchain and then put pressure on Steemit when necessary to make sure inputs from outside Steemit are properly considered when adding new features and making changes to blockchain economics.