Long term follow up. I eventually got to demo both the Mahr 817 CLM and the Mitutoyo LH-600E. TLDR is that both are very nice instruments, but for our use the Mahr edged out the Mitutoyo.
I'm going to ignore the pros and cons of the sales staff for this post as those are regional issues rather than technical differences between the devices:
I liked the Mitutoyo probe carrier a bit more. It had two positions to put the probe into, which allows a bit more height.
The Mitutoyo model had several external plastic bits which feel a bit cheap, but don't negatively impact functionality.
The shape of the Mitutoyo base was a bit easier to hold on to. My hands fit a bit better into the notches for moving it around.
I don't have a well enough temperature controlled space to distinguish these. On paper the MIT is slightly more accurate, but that's effectively in the noise.
It depends what you want to do, but IMO the Mitutoyo wins this one. It has two modes, although it's a bit of a pain to switch between them. There is the usual air float that makes it easier to move the gauge as if it was on an air hockey table. There is a second lower pressure/flowrate option where it takes some weight off, but doesn't really pick the stand up much. This makes it possible to use the glide function when measuring flatness (technically parallelism). I measured <0.001 mm rise when engaging/disengaging this function. On the Mahr stand the air glide is much stronger. It makes moving the stand much easier, almost to the point of being too easy. It takes a bit of practice to not have it move slightly when you engage/disengage it. While I like this for moving it around, it also lifts the stand several tenths when it engages depending on how you are holding it. This makes it harder to take measurements which require sliding the gauge. There is an adjustment for this somewhere under the battery, but I have not yet dug in to it.
The buttons on the Mahr stand are a bit easier to press, and more importantly it seemed to take a few more button presses to do things no the Mitutoyo than the Mahr. The buttons were also not as intuitive. On the Mahr after having the first two buttons explained I was able to easily figure out how to run the rest of the common functions without any additional explanation. On the Mitutoyo a bit more explanation was required, and I found myself having to think my way through how to do something a few times rather than having it be immediately apparent.
Both have this built in or as an option. I can see that it would be very helpful at times, but I have not yet tried it out.
While both gauges are capable of meeting their claimed tolerances on hole diameters, the fixturing required (specifically positioning or bore parallelism to the table) to get couple micron repeatability is not practical if other options are available. If you just need to be to the nearest thousandth it'll work great, but that is often more easily done with a pair of calipers.
The Mahr turns off a bit faster than I would like, and it needs to re-home the probe each time it turns on. AFAIK there isn't a setting to increase this timeout and it's somewhat annoying, especially with a somewhat involved setup. It also gets a bit confused if the calibration tool isn't at the height it expects. This is easily resolved by turning it on with a clear path to the table, but it's something I'd prefer not to deal with.
End of the day we opted to buy the Mahr due to the easier interface. I'll dig into the air bleed and see if I can get it to work a bit more like the Mitutoyo, but that seems easier to change than the software interface. If I did a lot of measurements where I need to sweep the probe over a surface I probably would have gone with the Mitutoyo as the two were close. When I measure something it's typically a 10 to 20 µm profile tolerance in an area that I cannot get a micrometer. I'll typically have 10 or so of these measurements, all of which are farther apart than the range of my Interapid mechanical indicator. In comparison to stacking gauge blocks it's a ton faster (no surprise), and this thing definitely pays for its self. Given the third alternative of readjusting the a height master and then adjusting an indicator to match, this is still much faster.