Yes, my pages - and as a result of this discussion I've extensively revised the 'Waypoints Limitation' page that you mention - and by the look of it, I may have to revise it some more ...
I opened your routepoints3120.gdb file and I see that you've just clicked randomly on the map.
Not quite like that. The first route of 160 routepoints (in file routepoints960.gdb) is a pukka Mapsource route composed entirely of Map Points.
All the remaining routes including the ones you looked at are just a copy of the first one with the coordinates displaced. They still seem to me to be behaving as Map Points because they still have all the road junction etc info embedded in, that the first version had.
According to Garmin support, there's some kind of "map matching" going on when you create routes in MapSource using place names which exist on the map. The upshot of this is that to maximise capacity on a Garmin mapping GPS it seems that routes must be created using MapSource, and in addition you must upload the map tiles containing your route to your GPS.
Yes on a map-enabled GPS there are thousands (millions?) of waypoints already loaded into the GPS, embedded in the map, known as Map Points. Using the Route Tool in Mapsource makes use of these wherever possible. However the resulting route still has to be able to work on a non-mapping GPS - the Route and Routpoints, once constructed, do not depend on the map. So it appears that (some of) the Map Point data is copied into the Routepoint. The map may give a richer navigation, but the route works (including detailed turn instructions) without it.
As you've found you can't do the capacity test by duplicating and re-naming a single route because the points are not all unique. If you upload duplicated routes you will be able to get the stated capacity.
Yes that was what occurred on my first attempt. After a bit of head-scratching, the 2nd go involved more processing of the duplicated points to make sure they were being seen as unique points. So this time a limit is eventually found, for the old Legend C somewhere over 2400 - I imagine the precise figure would vary depending on how info-rich the points are - how many chars in the road names etc.
Before I took my issue to Garmin, my real, intended route with no waypoints would consistently fail to upload to my GPS at a shade under 5000 via points. When they told me I must upload the matching map tiles, I found this increased the limit to almost 7500 via points.
That's a remarkable finding.
If this "map matching" (Garmin's words) works as they say it does it seems to me that it has consequences for all third party mapping users.
Yes indeed.
BTW, you did well to have such a meaningful dialogue with Garmin I must say!
Well please don't imagine I'm pretending to have all the answers here - oh no - but my observation is this:
Using the Route Tool in Mapsource, clicking on the map will:
a) utilise an existing User Waypoint, if it exists, copying much of the data into a Routepoint. This effectively uses two slots in the GPS memory, 1 for the User Waypoint and 1 for the Routepoint. (Though you always have the tickbox option not to upload your User Waypoints.)
b) failing that, utilise an existing POI (embedded in the map), again copying much of the data into a Routepoint. This only uses one slot of GPS memory, because the POI is embedded in the map.
c) failing that and most commonly, utilise an existing Map Point. That is, any road junction plus a whole lot of other on-road points in between. Again, data is copied into the Routepoint and again it only uses one memory slot in the GPS.
d) failing all the above, generate a User Waypoint and a co-incident Routepoint. This occurs if you go 'off-road'. It uses 2 slots in the GPS.
By this logic, using any 3rd-party mapping for planning is liable to result in condition (d) above occurring all the time, so any Route will contain duplicate User Waypoints and Routepoints.
The only such 3rd-party tool I've tried is Memory Map - and in fact with MM it is not so. The MM Route Tool generates Routepoints only, no User Waypoints. Presumably they are also quite info-poor so might be quite efficient in terms of memory usage - I don't know, haven't explored that far.
If you are interested, I have a .gdb file containing 45 routes totalling, at last revision, 7771 via points which cannot be uploaded to my GPS even with the correct map tiles on the unit. It's very detailed and has far more points per km than you suggest is necessary! I have managed to get the entire route on my GPS by taking 18 routes out, converting them to tracks and then uploading both routes and tracks. My guess is that a Legend HCx has the same basic capacity as my 60 CSx
Sounds huge! Please by all means send it - my email is in the sidebar (I don't seem to be able to remove it
(please zip it if poss - I'm rather averse to large files)
The Legend HCx and 60CSx have different chipsets - the 60 being generally considered to be marginally the better, I think.