NavList:

Matus Tejiscak, you wrote:
"It's not clear to me what the plan is. Assuming we fix UTC at a constant offset from TAI, what will the civil time be based on? UT1?"

No. The vast majority of practical time systems will be simple fixed offsets from UTC, just as they are today. Here in New England, civil time is exactly four hours offset from UTC for most of the year and exactly five hours offset from UTC for those months near the winter solstice (we switched two weeks ago). 

You added:
"I assume we don't want to decouple the civil time from the Sun."

Of course we do, and we have been doing so for centuries! When was the last time you encountered a train schedule announcing arrival and departure times in "sundial time" (or more techically "Local Apparent Time")? It's been two centuries since we kept time by the Sun.  By using mean time we broke away from Sun time by up to 16 minutes during the course of each year. And Sun Fast Day was also just two weeks ago when mean time clocks show their largest deviation from apparent time during the year. 

The trend away from Sun time continued. By the last nineteenth century it became clear that locally derived "mean times" were no longer useful so most countries and regions (not all) soon switched to zone time. At first this added up to about +/- 30 minutes additional deviation from Sun time. Within a few decades many regions realized it was more valuable match the clocks of important cities rather than be a bit closer to Sun time and they shifted from one time zone to another (usually eastbound) adding yet another half hour or more deviation from Sun time (example: in the US the western limit of Eastern Time used to be the railroad hub of Pittsburgh. Ohio and Michigan were on Central Time matching Chicago, the "Second City", but they legislated themselves into Eastern Time to match NYC in the early 20th century). And then of course there's the annual oscillation of "daylight time" or "summer time" or whatever it's called locally adding yet another hour. This implies that during most of the year, midwestern US cities like Detroit and Dayton and Indianapolis keep time as if their citizens are living in Barbados! The difference between Sun time and clock time or "civil time" can be quite substantial.

You wondered:
"Will we rework all time libraries to include DUT1 tables for conversions between UTC and local time?"

For those handful of applications that depend on "Earth orientation time" or UT1, yes, that's what's required. Astronomical observatories, for example, have sidereal tracking systems which are frequently legacy systems implying that the original programmers are long gone (sometimes literally). Such systems may include hardcoded one-seconds limits on DUT. If so, they will need to be fixed, either modified or re-written from scratch, in the next dozen years. This is not difficult, but it should be recognized as an expense. For celestial navigation, the issue is similar but much easier since there are very few legacy software systems. We can deal with large DUT values easily.

Frank Reed