bdiscoe has commented on the following diary entries

Post When Comment
New mapping addict 3 months ago

Hi Jimmy, welcome to OSM. I took a look at Launceston; it's a pretty large area. I started to work in the southern part, Prospect Vale and the area S of Meander Valley Road. Is there any particular part of the Launceston area that could use some attention from those of us who are advanced armchair mappers?

Come work on #MissingMaps with me! 4 months ago

Hi SDavies, you're probably right. E.g. I contributed to and rapidly found myself the only person doing that one, too.

BTW, I took a stab at Project 727 just now and filled in the last few tiles to get to 100%. Yay!

Come work on #MissingMaps with me! 4 months ago

Hi Warin, yes it seems that in many developing countries, even with a lot of tree cover, where the people live is very bare (possibly because of grazing animals, or firewood collection, or both). So, the roads (and paths, buildings etc.) are all very visible. You can see on that link (#793) that the target is a large town / small city, so it is not just roads.

Map Somalia 7 months ago

You will find that almost all of Somalia's current OSM data (especially northern part of the country) was uploaded in 2008 with the change comment:

"Mart Roumen; FAO-SWALIM; Madmappers-SOM-2008-01"

The good news is that because of that 2008 upload, Somalia has at least some data, it is not blank. The bad news is that the accuracy is very low. There are many problems, especially with classification, for example, many almost-nonexistent trails (temporary camel tracks?) are marked as highway=primary.

I spent a lot of time to clean up the "Mart Roumen" data in many Somali cities and villages (Borama, Hargeisa, Burao, Laasacanood, Berbera, Galkacyo, others). But of course, it is a very big country and I cannot possibly do it myself. Good luck!

JOSM scripting plugin: be a power user! about 1 year ago

For those interested, a later version of the script which handles "blvd" etc. and is better written is available at

The first 30-day challenge: retrospective about 1 year ago

@rickmastfan67 I'm not entered in the new contest. I was just testing out the freshly released, newly improved maproulette for curiosity and to support Martijn's great efforts. There was an initial glitch with Chrome not popping up a confirm on fix, which got fixed. I do wonder, though, how this new contest will work, since the maproulette issue aren't many, it seems they'll run out very quickly. Maybe Martijn has many more kinds of detectable issue coming soon.

The first 30-day challenge: retrospective about 1 year ago

@rickmastfan67 I am very impressed by your work on interchanges, which sounds much more laborious than general road cleanup, which would explain why it takes more time. It's valuable and appreciated! I wish i had an "auto roads" that worked, but that was just a programming experiment some months back. As for the term about "automatically [..] edit repeatedly", it doesn't seem useful to interpret that in a way that would prevent, for example, the use of JOSM's wonderful, powerful auto-validation commands. I certainly review every individual change before commit (as should everyone, contest or not) so that the changes are not "automatic", as a bot would do them. In any case, I had initially thought that fixing abbreviated street names would be a big help, but in the end it turned out to be a very small part of the changes I made. The great majority were: 1. Cleaning the noise out of really bad Tiger roads in South Carolina (JOSM Ctrl+F Find "way highway tiger:reviewed=no", then Shift+Y Simplify, then a quick manual touch-up pass over the affected roads.)
2. In endless suburban sprawl where the geometry isn't so bad, using Shift+T to add the missing turning_circles, scooting a few nodes to align with Bing, confirm the name and then T to clear Tiger.

The first 30-day challenge: retrospective about 1 year ago

@YourVillageMaps I wish I could record and share a video of me doing Tiger cleanup in JOSM with my shortcuts. The pixels fly! :) I also wish i could teach others how to do it. Telenav should hire me to train a building full of interns to do it, then they USA would be clean in no time :) (It also helps to have fast internet, a large monitor and a good mouse.) I am very proud of the work i got done in this contest, for example. the entire city of Lumberton, NC ( (And Dillon, SC, and several other towns in that region.) I also did tons of work like adding buildings, fixing streams and wetlands etc. even though they didn't count for the contest, just because i love editing. I also deleted lots of entirely non-existent Tiger roads, which also didn't count. I also fixed/added hundreds of bridges (over those corrected streams) which presumably counted as "adding a road", since bridges do involve inserting a way.

Making JOSM faster with javascript keyboard shortcuts about 1 year ago

Hi Theodin, Yes, some of the Tiger tags are useful up the point where the road is reviewed. At review/cleanup time, you often have to make sense of a mess of bad geometry, mis-naming, etc. and some of the "tiger:" tags can provide clues on how to make the road correct. Once it's correct, though, there really isn't anything in those tags that's relevant. And, even if someone wanted to scan obscure things like "tiger:tlid", for example a program doing some kind of mass-analysis on the history of the tiger import process, it's still available in the way's history. Nothing is lost in OSM :)

JOSM scripting plugin: be a power user! about 1 year ago

Woodpeck, this is no more of a mechanical edit than using JOSM's existing (powerful!) validation cleaning tools. For example, if you have isolated nodes, or duplicated nodes, JOSM can fix them for you in a single mouseclick. This script only provides a cleaning tool that JOSM hasn't implemented yet.

Of course, mechanical edits where visual confirmation would be needed, are bad. But, it does not take visual confirmation to rename "Main St" to "Main Street".

Project: Puerto Rico TIGER Fixup about 1 year ago

How do we contribute to the project, other than by simply picking a place in PR and start editing? I did some tiger cleaning in the town of Arecibo just now, leaving the tiger:reviewed=no tag on the streets where I couldn't tell if it's correct or not; those should be visited by someone on the ground or local knowledge.

You also mention "USACE 2009-10 imagery", how do we use that imagery layer?

Lost city in Darfur over 1 year ago

Equiserre, great work! I am glad to see you found names for places like Um Danga which I had mapped but left unnamed.

Alexz, I have now made a wikipedia article at

Lost city in Darfur over 1 year ago

Thanks everyone! Amazing research and maps you've all found. I've added an OSM place=town node to indicate that this is "Gereida", with "Gerida" as alt_name.

Lost city in Darfur over 1 year ago

Thanks alexz! I'm not sure if the big town I found is "Gereida", but the town farther south is definitely Buram, so I added Buram there and named it. As you say, Darfur has clearly changed a ton since those 1963 maps (I picked a well-known city like Al Fashir, and it's almost unrecognizable now from how it is on those maps.)

Auto roads, part 2 over 1 year ago

booo: Better than releasing the source code, if it actually works well, I'd want to see it as a standard feature in JOSM. But, I'm not assuming success here, anything resembling computer vision/AI tends to have mediocre solutions at best :)

Auto roads, part 2 over 1 year ago

Aury88: Right now the code assumes the road curves no more than +/- a few degrees per step, so it won't turn sharp corners; it should stop if it reaches a T junction. If two roads intersect at a very sharp angle (a narrow Y junction) it might find either of the roads going forward; this is unavoidable, even a human might not know which is the "correct" connecting road. As for rivers, it could easily follow very regular-looking streams which don't turn too sharply, but most rivers are too large or messy (I find that the existing JOSM plugin "FastDraw" already works well for them.) Building outlines are almost never clearly linear, so I would use a whole different approach for those (like the iD research guys have shown).

raykendo: Yeah, I've been thinking of doing that as an extra step. I'm already scanning forward to find which way is the most "similar"; a second scan could look for "middle", perhaps by looking for symmetry in color. Then we have two competing directions - "this way looks like the road continues", "this way looks like the middle of the road", and I could try blending the two directions, perhaps based on the confidence of each result.

First attempt at automatic road following over 1 year ago

That approach actually worked in many cases? I'm glad to hear it, and I look forward to trying it someday, but it seems like a no-go from the first step, flood-fill. Most roads vary considerably in color over the course, from one end to the other they may be wet or dry, shadowed or not, pavement or dirt conditions varying. A flood-fill would have to be loose enough to find the whole road, but not so loose that it spills far outside - and those bounds would have to be adjusted from place to place. How could that work?

Automated road tracing - "Microsoft Road Detect" didn't work for me over 1 year ago

Yeah, the second discussion there points to the JOSM plugin "scanaerial" which is very hacky, hard to install, and only really for water bodies not roads. (It's also temporarily broken at the moment; can only reach landsat not Bing.)

I'd be curious why that 'never-deployed Potlatch 2 branch" didn't happen, and if it was promising or not..

I'd seen Mapbox's teaser as well, which is clearly still pretty researchy.

I'm going to try my own, more in next post..