[dredmorbius]

What you've done above and beyond GTD as I noted in an earlier comment (<https://news.ycombinator.com/item?id=46376558>) is that you've time-blocked when you make that assessment, on the (IMO correct and insightful) notion that triage itself is expensive. That is, determining the importance and length of a task / item of email itself consumes limited cognitive resources.

Adding to what I wrote earlier, another advantage of postal mail is that it comes at fixed intervals, typically once a day (historically possibly more often especially in cities with "morning" and "afternoon" mail, making one-day responses possible, currently with curtailments in service possibly only a few days a week). This automatically batches mail processing.

Early in the corporate adoption of email a firm I worked at only polled periodically for new external email (every 20--30 minutes or so). Whilst internal email was pretty instant, this meant that at most external emails would give cause for interruption only a few times an hour, rather than at any given moment. I've given thought to reimplementing this on my own systems from time to time, perhaps even only 2--3 times a day, say, "morning email" (limited to priority recipients), and afternoon email (the Great Unwashed Masses have their opportunity).

In reality, I've adopted Inbox Black Hole, in which I rarely if ever check personal email. Circumstances make this reasonably viable, though those are decidedly atypical and most professionals would be unable to adopt a similar tactic.