I'd never considered a full-scale hexagonal progression, so I just gave that some thought. An interesting idea for sure, since it eliminates the hex vs. square transition. Self consistent, almost organic. It also strikes me a bit impractical, but I wanted to think about why it struck me that way.
It's similar to another idea I did think about, which was throwing the hex out altogether, and going full cartesian. You get the same consistency effect, by not mixing the styles of geometry. When we draw maps, we like to measure position and distance in the context of a cartesian coordinates.
The question here is, do you still use squares to approximate a location (just swap hexes for squares), or go with real coordinates? A coordinate star map is pretty cool looking, and adds an element of realism to the game, but not without cost. A player glancing at the map cannot determine viable jump routes without closer examination of the map or additional information.
Using square shaped spaces gives you simplified positioning, which makes movement easier to plot. Despite its natural fit in the cartesian grid, however, the square is a lousy movement grid. Great for position and distance.., but crappy for movement simulation. This is where the hex comes in.
The serious minis wargamer never needed any grid, because the tale is in the tape, so to speak. The tape measure tells you everything you need that isn't already obvious by looking at the table. Grids are used to abstract position and movement for games that can get by with simple counters -- much cheaper and easier than armies of painted miniatures! Orthagonal movement is awkward and unrealistic, if very easy. The hexagon therefore becomes a kind of compromise between gamism and realism. We give up easy cartesian positioning for improved movement simulation.
To me, though, going with a full-hilt hierarchical hexagon based system, is kind of magnifying the artifice beyond usefulness. Since movement is plotted only at the parsec level (and not at the subsector, sector, or domain level), its uselfulness on a Traveller star map stops there. IMHO, YMMV, etc.
Regarding cartographic and presentation issues...
IMTU, a hex represents an area location rougly 1 parsec in diameter, just as in the OTU. Subsectors should be large enough to encompass one or more small-to-medium sized clusters or pocket empires. It should also typically contain enough resources and wealth to justify a standard ducal fief (this can vary).
I am not plotting star positions on any strict coordinate system beyond placing them in a particular hex. Any particular jump rating is assumed to have the necessary range to reach destinations based on hex displacement only. No specific assumptions are made regarding the precise range of displacement delta for any jump rating.
Not straying very far into heresy so far.
However, I really do like Rob's idea of collapsing sector sizes for MTU, so my sectors will be 4 subsectors in area, with 4 sectors making a typical domain.
For presentation, I am targeting both software display (possibly web, rich .net client, or both) and letter sized printed pages. The printed page should fit a complete subsector within the upper margins of the page while maintaining roughly canonical hex sizes (or bigger) on the page. The remaining lower portion of the letter page will contain world summary data and notes. I like and intend to cram as much symbolic data as I can into the hexes, so slightly larger hexes than the typical TAS form are oki-doki. Less lists, more geometry!
A sector map can easily fit in this same area on a single page as a dot map. Domain maps are not a major concern to me.
Didn't mean to burden y'all with all that, but since it came up...