Re: [OSM-talk] Overlapping Crossing ways
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Richard Fairhurst wrote: Mark Williams wrote: [snip] On a slightly related issue, I'll be interested to see how changesets, and the easier ability to monitor an area, affect this. 0.6b went live on 15th January, 0.8a on 30th March, so these duplicated ways have been around there for a while: I'm slightly surprised that they hadn't been noticed before. Mmm, me too, but I've been looking to the east more of late, so it took 'til I ran up Validator over an area I knew to be reasonable already to spot; but actually the whole section of M25, A13, junction river were on twice plus, as I said, a couple of triplicates (!) This part has been done since before I joined OSM, so it's not under active editing (not by me, anyway!). Where it had been quite neatly done, it was hard to spot; and the tripled layer actually used the same nodes, so GOK Perhaps when routing is in actual common use these things will be spotted sooner. It would be possible to put in quite an exotic set of these, producing some really amusing navigation directions from that perspective ;) - I've had my Navteq system suggest the 7th exit from a 4-road roundabout before, perhaps I now know why? Anyhow, I'm only speculating as to causes; I have sent a couple of (hopefully) polite helpful postings to named individuals, but no feedback has been forthcoming as yet. Please don't take it as yet another dig at Potlatch - As I said, I think it's more of a network issue and would affect any on-line editor. Mark -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFIQX3CJfMmcSPNh94RAgwnAJ9acuTHcS4uewcfxenubt0DPX8A4ACfSs73 AZkhUK6ZdYdFT5Qf+TwonQ8= =dMrh -END PGP SIGNATURE- ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Overlapping Crossing ways
On Fri, May 30, 2008 at 9:13 AM, Richard Fairhurst [EMAIL PROTECTED] wrote: (Should I also just disable z11 on Potlatch altogether? Thoughts welcome: personally I don't use it, but then I don't use the Yahoo imagery either.) I'm an urban mapper, so the higher the zoom the better. Especially when clicking edit on any of my traces, which immediately tries loading tens of square kilometres of south-west london, bringing the app to a juddering halt. At least erring on the side of over-zoom allows people to easily zoom back out, where as starting at too low a zoom often makes the experience a bit naff. Cheers, Andy ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
[OSM-talk] Overlapping Crossing ways
This is another Potlatch grumble - but it's not really Potlatch, it's the internet. I just found a whole lot (really, a lot) of overlapping ways with some really old data (Potlatch alpha before) buried under new versions from Potlatch v0.6a 0.8 - for instance, the M25 Junction 30 had 3 (three!!) layers of roundabout in places. 2 hours has seen most of it off... I wonder how much of other peoples' time I just deleted? I can't imagine this was done knowingly; I suspect that the redraw was sufficiently slow that it looked unmapped to a less experienced user, and allowed time to re-draw the ways before it showed up. Would it be possible to make it draw ways nodes BEFORE the aerial photography, so this can't happen is less likely? I had assumed that having to move from the map to the editor would prevent this, but I think folk are scrolling miles never seeing the rendered version. Maybe a slippy map under the photography? Or make it impossible to add any nodes until it's completely finished loading? I also had a good number of "High Road; Low Road" names to untangle, and a river to unplait. I do hope it stays done this time :) Mark ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
Re: [OSM-talk] Overlapping Crossing ways
2008/5/30 Mark Williams [EMAIL PROTECTED]: This is another Potlatch grumble - but it's not really Potlatch, it's the internet. I just found a whole lot (really, a lot) of overlapping ways with some really old data (Potlatch alpha before) buried under new versions from Potlatch v0.6a 0.8 - for instance, the M25 Junction 30 had 3 (three!!) layers of roundabout in places. 2 hours has seen most of it off... I wonder how much of other peoples' time I just deleted? i had a similar problem near me recently. i had a lot of streets that had been 'helpfully' drawn, without tags, in small, unjoined segments. after i mapped the area in question, i merged all the relevant segments and tagged them as appropriate (all in potlatch, approx. version 0.9a i think). i was looking at the rendering the next day, and realised that when they'd been merged, a new long road had been created, but the original short pieces had not been deleted - not at all what i had imagined would happen, and surely not intentional. i guess this is what happened here with your m25 problems maybe we need an extra warning category in maplint, to indicate duplicate ways - i.e. using the same nodes and having the same tags? I can't imagine this was done knowingly; I suspect that the redraw was sufficiently slow that it looked unmapped to a less experienced user, and allowed time to re-draw the ways before it showed up. Would it be possible to make it draw ways nodes BEFORE the aerial photography, so this can't happen is less likely? I had assumed that having to move from the map to the editor would prevent this, but I think folk are scrolling miles never seeing the rendered version. Maybe a slippy map under the photography? Or make it impossible to add any nodes until it's completely finished loading? I also had a good number of High Road; Low Road names to untangle, and a river to unplait. I do hope it stays done this time :) ___ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk