So I came on board with the specific mission of developing re-usable HTML templates for our emails. What a lot of people outside of the email development world don't realize however is that creating really great, reusable, and scalable HTML templates is actually quite difficult and requires a lot of specialist knowledge. So I set out to find the best way to make sure that if I were to happen to be run over by a car one day, that my replacement would be able to quickly and easily get up to speed with maintaining and developing our template system. To that end, I settled on using MJML as a framework for all of our emails. I sold it chiefly by quite simply doing it - through community support and online help I familiarized myself with the language and used it to build our email templates. But further I was able to present it to the team in such a way that it was more about what the language can do for us. Quick and reliable modifications, more easily create new template layouts, and more importantly there's little to no 'lock in' for me being the principal developer (this is an important aspect to me and is one of my core values). So by pushing MJML forward myself, and then by demoing it to the team and showing (rather than telling) them the benefits, we've managed to adopted it 100% for our marketing emails and are in the process of updating triggered and transactional emails as well.
So I've been using MJML for a while now and thoroughly enjoy it. There's some differences here or there between the syntax but for my part I've found MJML to be considerably nicer to use and they seem to be innovating a little more quickly than Foundation as well.
Check my previous response for a little bit more detail but short story - there's a fair bit of personal preference involved and I personally enjoy MJML: https://litmus.com/community/discussions/5407-what-is-your-email-framework-of-choice#comment-8447
For past clients and for my current position I have built an "Email Archive" which houses a copy of each email as an HTML file which is linked with a simple tree nav that lets people find specific emails quickly (brand - category - date is our structure).
It depends on your ESP in a lot of ways. If your ESP can handle replies then it further depends on how it can handle replies. I've used a setup before where addresses that reply with common out of office content are put on hold for 30 days the first time, and if they are taken off hold and the reply comes back again then they are put on hold for a year, and if the reply comes back after that then they are unsubscribed entirely. If you're using an automation solution then I really, really wouldn't be handling these things manually because it will be a huge burden on you and really you probably could be spending time doing something else either more productive or more fun. If your ESP can't really apply rules to the replies, but it can handle them then I'd personally add them to some sort of suppression and deal with them in bulk later, perhaps a sort of soft re-engagement thing.
Seems they put it back...for now! I would absolutely love to have a contact at Gmail that could talk out the logic of these sorts of things.
It depends a lot on your goals and your audience and there's usually no 'silver bullet' formula but there are a few ways to look at it. Do you want to exclude someone who hasn't opened anything in the past year, or do you want to remove someone who's opened fewer than X percent in the past year? So rather than basing your exclusion on number of days since last engagement it may be useful to look at how many opens your most-engaged subscribers are giving you and then start an exclusion from there. Say if your top 5% of subscribers open 35% of your email then take people who open less than 25% of the mail you send them over the last year and exclude them from your list. This gives you a little flexibility in creating segments and also might help you create segments and targets for re-engagement campaigns.