NavList:

David C, you wrote:
"Transverse Mercator was being introduced as late as 2000."

Probably not, or not as late as it appears. First of all, this was probably a legacy data recovery or legacy data preservation project for maps that were drawn and printed decades earlier. I would bet that this project was originally funded and launched many years earlier, maybe as early as 1990. It might still have made some sense back then.

You quoted part of their explanation for the project:
"however it is a projection unique to New Zealand and so can be difficult to use or program into computer software or positioning devices (eg, GPS receivers)."

The money they spent on the conversion project was probably quite impressive. You can be sure it would have cost much less to pay the commercial produces of software and devices to include the old, unique NZ projection system. Coding up another algorithm isn't tough, and those companies thrived on that sort of thing twenty years ago. And of course, by the time the project was completed (or nearly so), it was probably already irrelevant. That legacy data should have been digitized straight to lat/lon points on the WGS84 ellipsoid. They could have done this. 

You added:
"It seems that Mercator is not dead in NZ!"

I realize you're just being contradictory for the sake of the discussion, but I need to say this: "Standard Mercator" is dead in NZ, and it wasn't an issue any way. Since New Zealand is not huge (I won't say small), and since it sits in middle latitude, the distortions in the Mercator "global" (near-global) projection were not a serious problem (though if you want, you can compare New Zealand to Svalbard on a standard Mercator projection of the world). The continuation of Transverse Mercator in minor applications is largely irrelevant.

I want to emphasize again that "Transverse Mercator" is radically different from normal, common Mercator projection in practical application and practical implication. For the regional mapping function you referenced in your post, these TM projections were almost certainly narrow, overlapping strip charts running north-south. That is, the new 'equator' in these TM strips, comparable mathematically to the real equator in standard Mercator, would have been meridians of longitude, and probably several meridians of longitude were employed. One can detect this by superimposing the overlap regions. They nearly match up, but not quite, so when you read off a point from a TM map, you need to determine which longitude was used for its centerline (pseudo-equator). 

Transverse Mercator will be around for a long time in specialty mapping applications serving minor interests. Again, this is really irrelevant to the issue of the global Mercator projection which was removed from Google Maps last week. Those TM projections are also completely irrelevant to the Mercator projection used on many nautical charts. They're not even remotely the same thing in any practical sense, and the Transverse Mercator projection loses essentially all of the familiar properties of ordinary Mercator charts. To reiterate, if you draw a straight line on a Transverse Mercator chart, it is not a rhumbline. That's a key feature for most folks who have learned about the Mercator projection in a navigation context. If you lose that, then what do you have left? ... apart from the defining property of local conformality?

Frank Reed