see here for some examples
http://wiki.openstreetmap.org/wiki/Key:tracktype

There's no hard and fast rule, and there can be some overlaps between roads and tracks, but I'd say;

tracks are usually unsealed or of lesser quality build to a road - only wide enough for one vehicle - often for agricultural use or for property access - won't have road markings - usually privately owned etc.

Roads will usually be sealed, will usually have some road markings or road signs, usually maintained by the council (except for some private residential roads).

I've had a quick look at the Bing imagery, and the tags for the ways - and it looks like the way leading north-west from that junction has a partially gravel and partially paved surface. Also see that there are no road markings at the junction (like there are with the road to the south-east). highway=track is probably reasonable, but perhaps a tracktype=grade1 or tracktype=grade2 would be appropriate.

"If it's the former I agree that remapping would be premature. If it's the latter, I can't really fault Rovastar, tbh."

Which is why I suggested that Rovastar actually tries to get in contact with the declined mappers to see what their intentions are, rather than just assumes that their data is "dead".

Umm, but they're not "dead users". They are merely users who haven't accepted *yet* and may well still do so. Users who have declined can also change their minds.

From those users' points of view, you are simply vandalising their contributions.

Have you actually tried to contact those users and see what their intentions are?

I would say it's best to wait and see. It may be that those users will have accepted by the time the change actually happens.

13's are a very common one to miss out - usually going either 9,11,15,17, or 9,11,11A,15,17 - although one house I've done has 11X instead of 11A.

But there's one which is just mad: 1,3,5,7,9,11A,9B,11,13,15

9A might had made sense instead of 11A - and it's not like it's new build either - it looks to date from about the same time as the original development.

One way for the highway=unclassified. One way for the highway=track. One way for the highway=footway, then a final way for the highway=track. Four ways in total. Tag them all with designation=public_footpath as well. Adapt as necessary to the situation you have. If you are confident with route relations, you could add all four into a route relation to say that this is "Public Footpath RS66".

Oh, and highway=path. This is one of the ones several people can't agree on. From what I can see, most people in the UK seem to use it to mean an indistinct weathered route where it isn't clear what the access restrictions are.

Some people, however, do use highway=path to mean anything which isn't a track - and they add access tags to 'turn it into' a footpath/bridleway etc.

highway=footway is the usual tag, not highway=foot.

Other tags which you may find useful are;
highway=bridleway - I tend to use these for bridleways which aren't wide enough to be a track
highway=track + tracktype=grade* where * is a number from 1 to 5 http://wiki.openstreetmap.org/wiki/Key:tracktype
highway=byway - not in very widespread use - but could be legitimate if you come across a public-byway or restricted-byway which isn't wide enough to be considered a track.

For my eTrex Vista, I bought a set of rechargable batteries. I'd definitely recommend getting a set. I can generally get a good few days' use out of the GPS before needing to recharge usually.

rnw - interesting - I'll have a look. As I say, I've stopped there several times and haven't seen anything obviously looking like somewhere where I'd be able to post a letter - either externally on the forecourt, or in the shop. Next time I have to post a letter I'll go into the Texaco shop and ask whether they have a post box.

I've looked around for SK9 9. The location given is "Texaco, Altrincham Road". OK, so that seems pretty explicit. But looking at the particular Texaco (with the shop now a Somerfield) reveals nothing obvious. There *is* a post-box within 100 yards, but that one is SK9 70. Presumably RM have removed the Texaco one (or I haven't looked hard enough).

Sometimes it seems to take quite a while to load - so it's blackness for a while - then the tiles start loading.

You say that it's for anyone wanting to make a change to make sure they are obeying the rules - but signing the CT allows OSMF to potentially make that very change by holding a vote of active contributors. As long as the vote passes with a 2/3 majority, all objects in the database would be relicenced - and the only condition is that it's "free and open" - note that attribution not necessarily a condition of this.
The way I've read this section of the CT (again, not a lawyer) - is that the onus is on the data contributor to ensure that all data contributed is capable of being relicenced and you agree to relicence under any other free and open licence - which could even go as far as PD or CC0 etc. OS OpenData is clearly not capable of being held in, say, a PD or CC0-like database.
The upshot, if I've read these correctly, is that attribution only imports or data tracing are not compatible with that particular clause of the CT.

Right, they are in decimal minutes, and not seconds of arc. The reason I can tell is because most of the coordinates are also defined in SI 1999/762 - which defines the border in the Dee & Severn Estuaries. If you convert the coords in SI1999/762 from OSBG36->WGS84 - then the points lie extremely close to those from the latest document with decimal minutes - but much further away with seconds of arc.

I've seen the document for the Welsh border - and it has many points with greater than 60" - too many for it to be an isolated mistake. Perhaps they're not seconds of arc, but decimal minutes? In which case, your coord is 51º32.53' N, 2º42.95' W

I think it should be possible to put the 60CSx in "Mass Storage Mode". With that, if you have a USB cable, you can get the tracks as GPX files straight off the data card and save them to your computer.
These should be fine to load into OSM without doing anything, but if you want to remove any rogue points you load the GPX files in Mapsource, you should be able to edit the tracks in there. Make sure you save as gpx again.

But marking crossings *really* aren't "overkill" or "unimportant tedious details". Road crossings come in many forms. Toucan crossings are a UK term, but the cyclemap originated here. Basically, it's any signal-controlled crossing where both cyclists and pedestrians are allowed. That definition can be used internationally. The tags we use are just shorthands. Using these shorthands gets round problems with language difficulties.

Most road crossings in the UK and worldwide do not permit cyclists to cross whilst riding - so marking what type of crossing is actually useful information. In some places like Germany, pedestrians are forbidden from crossing until the green man appears; if you have a piece of routing software, you could build extra time costs for a way in your routing graph if a certain type of crossing is present. I really reject the fact then that crossing type is unimportant.

Unless you have professional GPS surveying equipment, then there is little reason to trust a single trace over multiple others, especially if yours is significantly different to the others. Use all of the traces and find an "average position" which hopefully will remove significant bias such as; GPS 'multi-path' errors, poor signal strength, having to move out to pass a parked vehicle etc. etc.

Having said that, you may often find traces uploaded where the signal strength was obviously quite poor and the traces jump around seemingly randomly. It's fine to put less "weight" on these traces, but I still wouldn't completely ignore them.

If you make your GPX tracks identifiable, then they are downloaded as ordered points with timestamps, so tagging with a source date is not necessary.

Secondly, in most areas, the geometry of roads/other features will not have changed at all over the last several years, so it's not clear why traces from, say, 2 years ago, are any less reliable than those taken today.

Thirdly, in many areas, there are multiple traces, and most people will use some degree of "averaging out" of GPS errors in getting the alignment. These traces could be many months or even years apart - so tagging source dates are either wrong - or will require many dates to be added.

"I also suggest that you need to at least discuss within the group the implications of Executive Order No. 45 or the PPCS-TM/PRS92 as a WGS84 variant"

PRS92 and the others mentioned are very localised datums (i.e. the Philippines), which will surely be inappropriate in Europe or the Americas. OSM is a global project and the datum used (WGS84) seems a reasonable choice, given that other choices could be inappropriate elsewhere.

landuse=grass should render

Hold on, that's not quite what I said. If you are going to actually mark them, then if it's effectively an internal note to other mappers - then the tag semantics are probably not important. Anything will probably do, as long as it's obvious what is meant by it. A note=* tag on an otherwise untagged way would suffice, where * = some description which says, in plain English, what this way is doing here, and that mappers shouldn't waste their time by going out and mapping it because nothing exists etc.

There are plenty of objects all over the globe tagged with note=* intended as notes to other mappers - usually on objects that do actually exist - but perhaps not all information has been recorded, or something is incomplete etc. You probably don't need to overcomplicate things by defining a complex system of tags to describe that something *isn't there*.