This relation is probably the culprit http://www.openstreetmap.org/browse/relation/74769.

This relation is probably the culprit http://www.openstreetmap.org/browse/relation/74769.

But now we have two boundaries for the Peak District NP.

Many National Parks in the UK have been mapped tentatively (see UK National Parks: mainly because of uncertainty about which data sources are not tainted by OSGB copyright issues. Many were mapped using relations: from time to time these get accidentally broken, and need to be fixed. I think something like that happened to the Peak District.

There are more on the wiki here: Routing/Taxi to

I leave to others as to whether such an import is a good idea, but the most widespread tag for tagging type of woodland is wood=deciduous|coniferous|mixed, rather than the suggested trees (for wood see taginfo, tagwatch etc; for trees see taginfo). Given that a large number of CLC polygons will be Holm Oak (Quercus ilex, encina) and Cork Oak (Quercus suber, alcornoque) it might be a good time to get wood=broadleaved (I don't think the hyphen is necessary) used as a common tag instead of the irritating misleading deciduous tag. It would then be easier to distinguish between evergreen broadleaved woods and deciduous coniferous ones, although another tag might be needed, such as evergreen=yes|no.

Well done: a year's worth of monitoring and mapping comes to fruition!

It looks as if most of the area was mapped from NPE data: GPS tracks are really needed!

The Summit HC does not have routing capability, AFAIK the routing information is not transferred to the internal memory of the GPS. You don't say which OSM-derived Garmin map you are using: there are several available for most European countries.

The styling on the device will be determined by whether you have downloaded a .TYP file or not. I have no problem with Mapsource compatible maps on this GPS from any of the usual sources (All-in-One, Computerteddy, OpenMtbMap, Lambertus, alternativaslibras.es or talkytoaster). Additionally I build my own, particularly transparent overlays, using mkgmap and GroundTruth.

For the small on-board memory of the Sumimit HC, it is most useful to have quite small .IMG tiles (2-3 Mbyte) as this gives more flexibility (for instance if going on a long journey).

NHD data is imported basin or sub-basin. The import areas are chosen by individual mappers according to personal interest. You can see who might be importing areas in Colorado, here for east of the Continental Divide, and here for west of the Continental Divide. Osmarender gives a reasonable overview of whether hydrography has been covered in a region (e.g., for Colorado.

Importing the data is only the first step, each import requires substantial effort to clean-up the data. I think anyone who has looked at Colorado basins is not keen on the idea of doing huge imports over Denver, Colorado Springs and other conurbations, as these may hamper the efforts of other mappers. Even sub-basins are pretty large and likely spill out into densely mapped residential areas. Furthermore if any water features have been mapped then it is an important courtesy incumbent on the importer to preserve any prior manual mapping work.

The French CORINE data import only imported areas where they did not interfere with existing mapping. I think a similar approach has been discussed for NHD data on the us-talk list.

HTH

Please dont move mapped objects on the basis of imagery which may have residual errors from the rectification process. Instead move the imagery (shift-drag in Potlatch, WMS icon in JOSM) to correspond with already mapped objects: at the very least they will retain their correct topology and remain relatively accurate.

Continual movement of map data will only degrade accuracy: particularly if the most accurate data (imported pre-quake) are moved to align with less accurately calibrated images. In the absence of control data, such as GPS tracks, there are limits to the accuracy of rectification. Furthermore we now have a huge range of different sources each with its own imperfections, and none have been corrected with on-the-ground data.

These are created every 5 minutes at : http://labs.geofabrik.de/haiti/.

I use GroundTruth with a simple rules file to build a transparent overlay of unnamed roads (unclassified & residential only). The rules look something like this:

|-
| ResidentialRoadHighZoomNN
| {{tag|highway|residential}} and not {{tag|name}}
| 22
|
|
| "NO NAME"
| {{GroundTruthLineColor|#ff0011}}
| 3
|
|-

I use MapSource to upload my maps.

There are a range of ways tried for mapping golf courses, the issue is that they are very unlikely to appear on the default rendering. Some sort of consensus on basic tagging exists, and of course don't tag for the renderer (e.g., natural=beach for a bunker).

Have a look at the wiki: Golf_course, and Richard Weait's blog entry on the subject.

tracks=* is used over 4500 times in Europe, so I think describing it as abandoned is a little harsh. It is very useful when it is too difficult or time consuming to map multi-track railway lines, and conversely tracks=1 removes any ambiguity for single track lines which have been surveyed in detail. For instance I am have recently been using it for lines of tramway where I have not split the tracks (necessary for getting tram stop access right.

My understanding is that one is a district and the other a London Borough. A quick google gives this link, which might be of some use.

RichardB, not all of us can match your herculean efforts mapping whole towns at a sweep (see ), and not everyone has a GPS.

Realistically mapping parties won't happen until the Spring, and I think it would still be a challenge to hit all 4 areas (+possibly Tameside) next year.

Either way an indication of housing types etc. on the wiki is an assistance for planning any kind of mapping.

Here's my suggestion.

I would imagine there are still large areas of Bolton with roads and houses which exist on the New Popular Edition. For armchair mappers it's impossible to know where these might be. If you are visiting the area for mapping or otherwise it might be possible to annotate an existing OSM map with approximate dates and types of housing, and stick on the Bolton wiki page. Any areas which are predominantly pre-war can then be mapped from mapping by anyone who wants to. Local mappers can then focus on a smaller part of the whole area, although road names would still be missing. To do this you may need to map a few of the distributor roads in residential areas. In a way this is dividing it into Cake Slices.

Unfortunately lots of the towns in the vicinity of Manchester seem to be mapping blackspots. Bolton is not the worst: have a look at Oldham.

HTH and Best of luck.

many interesting points. Here's a few of my thoughts:

Definitive map can't be the database because of these legal requirements (from the Ramblers website): "Definitive maps are public documents and must be available for the public to view at surveying authorities' offices during normal office hours. There is no charge for viewing a definitive map, but authorities may charge a reasonable amount for photocopying. Some authorities have created online versions of their definitive map." Most of the one's I've seen have been hardcopy print-out from large scale OSGB mapping.

More interesting for us are the statements. These often contain lots of information and very little OSGB stuff, usually just Grid References. If these could be released under suitable licensing terms, perhaps with the GRs 'redacted' it would be very helpful.

NPE paths and tracks are more useful than not. Clearly around towns they will have been obliterated by housing, but more often than not there will be little alleys/ginnels/twitchells (select local term of choice) which maintain the rough line of the old path. Where I've mapped NPE paths I usually don't connect them to other ways, but when I pass a footpath/bridleway sign and one is marked then I modify the original NPE one accordingly. Even better I walk some of them in which case I replace the NPE one entirely.

Even older maps are useful. The public footpath represented by a short footway and Lenton Road here> is under threat as the private estate which it runs through have gated the W end. However a map of 1835 (Sanderson's 20 Miles around Mansfield) shows this route as a road.

OSM is surely mainly about mapping what's there. Sorting out the legal status of paths should be a secondary objective. In cities like Nottingham which didn't have to maintain a definitive map, and are only starting to notify PROWs now under

I'd always be cautious about borders on or close to rivers. Usually the border will have been statically defined but the river will keep changing. It's best to have two separate ways preferably not sharing nodes. If the border was a relation then adding the river as a member is a reasonable approach.

I presume the tracing is done in Potlatch: I've never found it that easy to create two ways very close (or overlying each other) in Potlatch. I can think of two approaches using Potlatch: 1) lay out the way with nodes which are not quite in the right place: later drag them to their correct positions (the ways are not then joined); 2) split the current boundary way and delete the portion which is in the way - draw the river - use history to revert to the way you deleted. If you use JOSM you can copy a portion of the boundary, paste it into a new layer, realign it properly, change its tags and then merge the layers. This is appropriate if the river and boundary are pretty closely aligned. HTH.