I can't believe I didn't stumble on this before. A .shp file with all the trails of Yosemite can be found here:
http://science.nature.nps.gov/nrdata/datastore.cfm?ID=47797

Thanks for letting me know!

My personal take on this is slightly different:
- Obviously, adding a new WAY without first checking if it's already there is bad. Not sure if this is what you were complaining about?

- But uploading a GPX track for a given area, even if it's already there, is very positive. This is especially so for mountainous regions where GPS is less accurate, but it's true in general. It allows one to make judgments about the accuracy of data. Mass data storage is cheap. A bit of redundancy in GPX data won't hurt.

E.g. sections of the Tahoe Rim trail have been uploaded 5 times or more. I love it!

- I've move literally tens of thousands of edits in cities all over where I've never been: the US is a big place and there are uncountable number of cities that have never been touched. We're lucky to have TIGER data, but the data is only mostly correct from a topological graph point of view. So there is a lot of value in selecting a city and starting to align this graph to the situation on the ground. It's very rare to stumble on places where edits have already been made and it's usually obvious also, so that's something where I tend to keep my hands off.

Note that the US sectio nof the wiki specifically encourages to align TIGER data, even for places where you've never been.

A nice example is the city of Fresno, CA: pretty large and almost completely unedited just a few months ago. It took me many weeks to align thousands of streets correctly, but the result is a much higher quality map.

I don't know if I've deleted some other smallish edits in the process, but if so, then that's too bad. A little bit of collateral damage is a price to pay in a cooperative project.

I've seen it happen with some of my own edits. Usually I send a friendly message to the one who did and things can be resolved easily. (They're almost always enthusiastic beginners who don't quite know yet what to do.)

Tom

> A path is a single groove worn by lots of walkers, in general you don't want to walk along them as they will often be muddy or difficult to pass. You should look to using highway=bridleway, cycleway or footway before even thinking about highway=path. You might be interested in the Style Editor.

This is absolutely not obvious from the Wiki Features page and from the discussions in the forums.

A footway, as shown in the wiki, is more like a 'road' but for pedestrians only. Usually in an urban setting and usually paved. It would be wrong to tag a small mountain trail as a footway: you're basically classifying entirely different things under the same name.

The cycleway on the wiki is the same story as footway. Something that's rideable by a regular bicycle.

The bridle way on the wiki looks very much like a fire road. Much wider. Note that it also says that it's equivalent to 'path + horse=designated', which suggests that marking it as such is equally valid.

Basically, the tags that you list suggest a structured, planned, usually urban, often paved ways. And, in fact, your message confirmed that the trails are correctly marked as path: a muddly groove that's often difficult to pass in park. Yeah, that's exactly how I wanted them marked. It's what mountain bikers are looking for, except that they're not muddy in California. :-) )

And I'ved looked at the style editor (and filed a bug request with Cloudmade): the best it can do, is mark show the paths, but there's no option to also render its name. Which makes it pretty much useless for hiking/biking purposes.

The Osmdiff example of Hessen is pretty awesome. :-)

I'm not sure if it would be immediately useful though: I tend to have pet peeve project (e.g. correctly align all TIGER data for the city of Fresno) then switch to adding all fire roads of Mt. Diablo Park, then add all new streets in some backwater town where I've never been. My edits are all over the state of CA...

But it could be a first start to at least monitor some places of interest. I've been setting up a Postgis database with OSM data for a different project. Will take a look at Osmdiff.

As for the comment by Philip: I have the impression that the OSM main page renders pretty much everything, including those almost useless administrative boundaries that were imported for the US. Indeed, I prefer that much more over the arbitrary non-rendering of data that's really useful for some people.

(Yeah, I used "it's" twice where it should have been "its". Should reread more carefully before hitting 'publish'.)

This guy may be from California, but he totally agrees nonetheless.

I know that's how Mapnik is configured. I think it's configured badly...

Oh, man, you have no idea... Looks like you're in my neck of the woods. I've aligned thousands and thousands of street ever since I discovered OSM sometime in November last year.

Tom

Ah, ok. In my case, those tracks are definitely not residential. In the US, the problem is usually just the opposite: that vast majority of data that was imported from the US census database has been marked as residential, even if it's smack in the middle of the Arizonan desert with no house in sight. :-)

Why would it be inappropriate to label them track? That's basically what I do for all unpaved fire roads on which I mountain bike.

I've wondered in the past whether or not doing just that is actually a copyright violation? After all, if we're not allowed to look at existing maps to add new stuff, then taking a picture of that map would be a really easy work-around. :-)

So, basically, I limit myself to only naming those trails for which I'm absolutely sure, either by actually going there (and taking pictures) or, rarely, because a friend who's familiar with the park tells me the name.

Copyright law is a real pain and I prefer to err on the safe side...

All the tracks on the map that are marked residential is because the data originates from TIGER, the free US census database that contains pretty much all the streets in the US. There are many problems with this data base, most important the fact that coordinates of roads are sometimes horribly wrong (though the interconnection graph is usually correct.) Another one is that it doesn't classify roads in a detailed way.

So everything becomes residential by default, until someone like you changes it. :-)

I just checked it out: the road has been tagged both name:D850 and ref:D850. That should be sufficient, right?

I assume the combination of tertiary and unpaved is what's doing it in...

Mac, Firefox, Flash 10.0r12.

CapsLock seems to work fine on Safari.

Tom

Agreed. The most common TIGER mistakes are cases where the general topology is correct, but where the placement is totally off (sometimes by half a mile or more.) But there are plenty of cases where it shows streets that don't exist (any more?) or vice versa.

Sometimes it's so obvious that it's safe to correct even if you've never been there. But if you actually live in that place, don't doubt for a second and go for it.

Wrt census data not being correct: this is really the norm rather than the exception. My general experience is that the network graph of roads and intersections is usually correct, but their locations can be way off. Also, the lower the density of a network, the high the error of the location.

E.g. dense suburban street networks can usually pretty good. Rural roads are often a disaster.

I discovered that website yesterday! Very interesting if we could use his data. I was thinking about emailing him too... Right now, I take pictures are every trail intersection to help remember those that I forgot, but the eastern part of the part have not seen a lot of exploration.

I believe there's still a backcountry weekend. I've never participated though...

I've been there many times. The fire of last year was not a Yellowstone event: only a small part of the park was affected. The kind of vegetation in Coe is relatively benign wrt fire and recovers quickly. When you ride though the area, you encounter small stretches of burnt trees, but you're never in long stretches of devastation.

Two months ago, there was a controlled fire in the more southern parts, along Coit road. I was out there: we saw huge flames a mile or so away and rode past section where trees were still burning. I'm pretty sure that just a couple of month into the spring, you'd have a hard time knowing it was there.

Here are some pictures of our ride during the controlled burn:
http://www.mtbguru.com/trip/show/7963-coe-ride-during-controlled-burn-in-cross-canyon-area

Works nice on my MacMini, except that the mouse scroll wheel doesn't work

It seems to work fine on my Mac.

Great work, BTW! Do you plan to create a open source widget to display this or is this for your own commercial project?