Oh, and you can also edit the wiki (wiki.openstreetmap.org) for more permanent things. You have to register separately for that.

Not really - it's mainly intended to blog you mapping activity.

The mailing list at http://lists.openstreetmap.org/listinfo/talk is the main forum for discussion (it's a mailing list not an online forum). There are various other lists at http://lists.openstreetmap.org/ for specific topics/locations.

EPSG:4326 is the default in JOSM for some peculiar reason, so you probably didn't change it, it was always like that.

Of course, if you change projection in JOSM the ones you've already done will now look distorted in the same way as they do on the slippy maps. If you walk around a building, do your GPS traces not appear with right-angles when loaded in JOSM? They should, because your GPS is just recording lat/lon and there is no projection involved at that stage.

Mercator preserves angles so a right angle on the ground should look right-angular in the Mercator projection, which is the one all our slippy maps are done in (though there's no reason they have to be, but it is computationally convenient and works well at large scales for exactly the reasons you;'re discovering).

At smaller scales, there is no right answer because it is all about representing a spherical surface on a flat plane and there are many ways of doing that.

Hi Simon, welcome. You're a bit on the edge of the area, but you might want to join the talk-gb-midanglia mailing list at http://lists.openstreetmap.org/listinfo/talk-gb-midanglia which is loosely centered on Cambridge. In particular we'll want to address completing Bury St Edmunds soon, and I've no doubt there will be interest in doing Sudbury at some point.

David

The person who would know about Guilden Morden is David Rusling osm.org/user/David%20Rusling - he hasn't signed it off on the Cambridge wiki page.

Hi Ickogg,

Welcome.

The Roman Road so far is largely mapped from accurate (hopefully) junctions with other roads and e.g. the paths linking with Wandlebury, with the intermediate bits done from NPE maps (which is much less accurate; helped by the fact it is straight though). Your traces for the intermediate sections are likely to be more accurate on the whole. It kind of peters out near Horseheath - when I was mapping the nearby villages I clocked the point it crossed the Streetly End road, but I didn't follow the track more than a few yards, so anything you can do to extend it that end would be very welcome. I noticed it changes status at various points: at some points it is a byway, at others bridleway or just footpath.

Are you local to the area? If so I'll contact you separately to introduce you to some of the other people working in the area - there's a lot of current work going on with footpaths and drove roads, making the detail in South Cambs really very high indeed.

David

Hi Paul, welcome. Yes, you're right, South Cambridgeshire is largely done to street level and it's several villages west of the A1198 Royston to Huntingdon road that are outstanding. It's me who has done most of the street work in Cambridge and South Cambs over the last two years, and there's quite a few people also actively mapping in the area and there's been quite a lot of work recently on footpaths and byways - and the level of detail is increasing dramatically now.

Sorry about the flag on Little Shelford - I was introducing a friend to mapping there (a long time ago) and he never finished it, though I did then fill it in myself and forgot to update the wiki page.

We recently got together to do a day in Bury St Edmunds and have about 70% of it done. I'm working systematically on Huntingdon at present. St Ives is now getting some attention. But St Neots isn't really on anyone's radar AFAICS so far. With Eaton Socon it's getting on for 40,000 people, and we usually reckon about an hour surveying and an hour on the computer per 1,000 people for systematic mapping, so it's likely to be a commitment of about 80 hours if you end up doing all of it - keep you busy for a while! The more you can make clear what's been done the better - perhaps set up a wiki page for it and divide it up into manageable sectors first?

Please do get in touch if you want any help or advice. We had a small Cambridge get together in the summer. Perhaps I should organise another one nearer Christmas.

David

Indeed. As well as the kind of detail that Mark is doing, there are two huge projects for Cambridge that we might want to consider: (a) shops - at present we only have supermarkets, post offices, pharmacies and key convenience stores. We could map every shop. (b) house numbers - we might want to adopt the German scheme for doing this, or consider our own. Perhaps we should update the Cambridge page with what needs doing, big projects and small. There are also areas that will need revisiting (not to mention the huge developments due in the next few years) - Arbury Park (now named Orchard Park apparently) is being looked after, but the area between the University Press and the railway is being developed now. There are new streets in the Ramsden Square area.

Are you going to tell us which town (in which country) that might be...?

The namefinder searches do accept postcodes, but the way they work is to try to find an address by a web search and then look up the address in our database. That's because postcodes generally weren't - still aren't - stored in OSM.

However, if this is becoming more common, I can modify the search to include postcodes in the index (as well as the translation type search). It won't happen until I reload the index from scratch which takes many days; I am tentatively thinking about doing this over Christmas. I will also invite other suggestions of things to include in the index which aren't at present before I do that.

It wasn't always so. Names (and icons) on areas are quite new. Mapnik only handles it "better" in the sense that it has collision detection for labels, and for many areas the node and "centre" of the area overlap. If the area is a strange shape or the node is offset, Mapnik will put the similar labels in too.

So I think the University nodes can largely be removed as they are redundant. They always were tagging for rendering. Where the corresponding area is missing the name tags, they need to be added.

A related issue is school tags. Throughout Cambridgeshire I've marked amenity=school for the area of the school grounds. But the buildings are usually near the road side, so I've put a node with the name to stand for that. If we put the name on the area too it gets rendered twice. The answer is, of course, to do the buildings as buildings, like the colleges have been. The problem is that this only really works in Yahoo covered areas, and this only extends a few km outside the city (elsehwere it often possible to get frontages, but the depth and shape of the buildings away from the road is hard). Even so, though, we get a label for the building and a label for the area. Sometimes this is a good thing - individual buildings sometimes have individual names within an overall college. But that's not so often the case with schools.

Re the college/university distimnction, you'll have seen that I originally put names e.g. "Pembroke College (University of Cambridge)" or "... (U of Cambridge)". If we put something like affiliation="University of Cambridge", or created a relation grouping the institutions, I guess that would work, but I'm not sure it would lead to a terribly helpful rendering (yes, I know we shouldn't tag for the renderer specifically). If you were doing a map specifically of University premises, it would be very useful (except you'd probably want to distinguish between colleges and departments), but in a general map, knowing that this group of university buildings is U of C and the others are ARU is surely helpful. So I don't know. Maybe both.

The detail in Cambridge City Centre is getting phenomenal, which is great. I do worry that when one goes more than a couple of km away from Reality Checkpoint, Cambridge and South Cambs villages aren't getting much attention over the original "named streets and POIs" that I originally did - thout Tom Judd is doing great work filling in the droves and dykes north east of the city, and several people have been working hard on footpath and bridleways in the area. I suppose it is inevitable that the ease of the aerial photography, the "sexiness" of the city centre over suburbs and it's where more people live or know means this will be the case.

David

Very useful I agree - I use mine all the time for mapping. But even better, you could put it on the train or the bus and help the environment.

I've been following with interest your progress on the footways in our area (it's me who's done the majority of the road mapping for South Cambs villages and Cambridge City; Itoworld's new tool makes watching for changes much easier). Donald Allwright has been doing more footways north of Cambridge, David Rusling out towards Guilden Morden, Tom Judd the drove roads and paths out in the fens near Burwell, so there's quite a team working on the paths around Cambridge.

You musn't use printed maps to get street names as that would (nearly always) infringe copyright, and defeats the point of making maps with OSM. You need to get street names by survey. If you miss one and can't re-visit, you can tag it with name "FIXME" to amke it obvious it needs fixing.

In the UK, each district/unitary authority has an officer who is responsible for naming streets. I made contact with the Cambridge person who has agreed to alert me when a street gets named, so I can make a survey visit.

I was contacted a while back re Arbury Park: osm.org/?lat=52.23354&lon=0.12019&zoom=16&layers=0B0FTF because new residents were having problems directing visitors, so I'm visiting every few months to keep this large development area up to date. Google's at least 2 years out of date.

As I go round, I've been noting where there are sites in development so I can visit again - but often they take a year or more to complete.

If you go to www.informationfreeway.org, zoom to level 12, and CTRL+CLICK on the tiles in the area you have changed, you get a higher priority manual request for that area. My changes yesterday over 6 tiles rendered within a couple of hours.

If waterways have been traced from NPE out-of-copyright maps they can lose 200m accuracy over about 3 or 4km (at least in my eastern part of the UK - I found them better in mid Wales for example), and the lines on them are so wide you couldn't get it accurately to get a bridge say. But trunk roads shouldn't have come from NPE, and as a canal is always accessible, there shouldn't be any problem improviong its accuaracy - except in deep cuttings or lots of overhangling wooded bits that are common on canals where the GPS signal drops out.

You might be interested to see how I've evolved my mapping technique, not least to speed up what I can do. See osm.wiki/index.php/User:David.earl
I'd definitely recommend using voice notes rather than pen and paper.

You can find out who created a Way in JOSM (and it can't be anonynmous because Potlatch doesn't allow anonymous users) - download the offending area, select an offending way and the user name shows up in the appropriate panel. You can then contact the user through the messaging facility at osm.org/user/ (remembering to replace anty spaces in the name with %20)

I had the same problem with battery life on my Nokia N810. I bought one of these: http://www.addonsworld.co.uk/product.php/36108/0 - only 6 pounds and works with AA NiMH rechargeables. It is the same connection on the N95 I believe.

I suspect if you're keeping the device in your pocket, your body is shielding the signal to some extent.

David