Since I started this thread, I feel obliged to update my current thoughts.
I championed the method of 1.5 diagonals, and I also saw merit in the '1st diagonal is 2, the rest are 1' houserule, but there is one important failing of both of these.
I agree with Shoe that, at the beginning of a model's move, it should be able to move into any adjacent square, even diagonal, for just one 'movement point' in order to keep movement in line with reach and area effects. If the model can do that at the first square, why not the second or subsequent?
I also see the merit in having the range rules and movement rules as homogenous as possible.
Thirdly, I appreciate the recent argument that there's so many other abstractions going on in the game, why should the movement rules be any different? Movement already has abstractions, like square base occupation (2X2, 3X3, etc), initiative order, etc. In light of all this, enforcing a euclidean diagonal seems excessive.
So, what's my point? Well, I'm starting to see the benefits of 1:1 movement. I continue to assess myself as a gamist, not a simulationist, but I also recognize that I'm wired to seek the simulation aspect if it also works. That part of me is still having a hard time dealing with the Far Realms diagonal influence, but I think I can get past it. Looking at my DDM maps and dungeon tiles, I think the opportunities for diagonal warping will be relatively light.
Also, I think that 'first diagonal is 2, rest 1' is not a good solution and is worse than 1-2-1-2 because at least with the latter the 1st diagonal is 1, which I see as crucial for balance.
Finally, I think that all gates to the Far Realms will be placed at diagonals on my maps!