Way 615565842: you were not entirely right in changing this road to unpaved: only 1 or two sections in the centre are dirt. In many places I can clearly see the centre line on asphalt surface.

Way 615602462: unless your imagery shows otherwise (and it may well, if it is newer) the centre section of this road is unpaved.

Way 611647522 looks really good. Well done!

Way 625669242 should be a dirt track, dto. the pre-existing unclassified road over the bridge.

Way 609334359 looks paved to me. All roads drawn by FB are offset in this area by 9 and 4 m.

Way 628067406: the cross roads in the far NW need re-shaping, as the road makes a sharp turn here and the other two streets linking with it do not actually go straight through. The section E of road สก.3011 is unpaved.

Way: 627882602 looks like a track to me.

Way 604751435 is the continuation of the unclassified road to the S and unpaved. As it connects two unclassified roads it should be the same classification.

Way 605388408: N part is dirt.
Way 604751435 is the continuation and also unpaved. As it connects two unclassified roads it should be the same classification.

Please ignore my comment above about way 605316314, my mistake.

Ways 605335362 & 605316314 are mostly unpaved.

The imagery offset appears to be > 7m, so all roads drawn seem to be out by that much. Having said that, there aren't many traces in the centre of the mapped area. I would revisit it, starting with the town near the lower right corner.

Way 626729198: on DGS this looks like a dirt road, but on Bing the E section appears paved and there is a lot of dirts spilled over the sides, so I think this may be paved now. I would be happy to leave this as it is.

BUT: go a few hundred m W to the village of Chiang Yuen. You can see that almost all roads (other than FB created) are offset. There is a major road junction and this is where you will usually find GPS traces. From that you can see that the offset is about 16m. This is significant. When you see lots of existing roads like this it's usually a good indication that somebody has been able to correct the offset before drawing roads. You should always tro to check the offset against GPS traces before starting to draw.

osm.org/way/624543246: most of this is a 2-lane asphalt road, with the centre line clearly visible, except for the SW end, which is a track. @ 14.0809452, 101.1350116 you can clearly see the paved road continuing SE (currently marked residential).

*** Most of the ways branching off this, however, are unpaved. ***

Offset seems ok, going by a GPS track a couple of km W.

The irrigation canal running parallel is offset.

About the middle of the long straight one of your colleagues has nicely drawn a couple of bridges. Unfortunately, there aren't any, the canal goes in a culvert, most likely not only under the roads, but also under or through the water body that the road crosses. This, however, clearly is a bridge.

Way: 616670002: you changed this from unclassified to service. On both DGP and Bing (slightly better) I can see 3 distinct road surfaces: wide asphalt with marker lines at the N end, then concrete, then dirt. Given that this services several buildings plus fields I suggest marking this as residential and splitting it into a paved an unpaved section. The continuation S over the bridge should also be changed to unpaved residential. The junction with the highway needs to be moved to align with the image (it's not even parallel.) The same goes for the other junction in the S. There doesn't appear to be an offset with either images, so the alignment should be good. That bridge was well spotted!

The second problem I flagged above: West end of Way: 629622938: I may inadvertently have caused this when I moved the whole area to account for the imagery offset. I'm sorry about that.

osm.org/way/629622905: See how you have connected the street to the one-way section of the highway? https://www.dropbox.com/s/s83jex32ya2ave0/OSMway629622905.jpg?dl=0
This prevents routing a right turn out of the street, as the routing software will not route against the one-way street. Trying to turn right into the street you would get a confusing instruction to make a U-turn, then immediately turn left, when in reality only a right turn is required. You can again see an imagery offset. Once corrected, the two nodes at the junction are on top of each other, so it's safe to just merge them.

To contrast this, check out the two roads branching off the hway N of this. The median barrier wall extends past both junctions, so it is physically impossible to turn R here. (without the drawn wall there would be no way of telling, because the median barrier doesn't show up in the images. Given this is drawn as a dual carriageway I would guess that you can't turn here and still draw it like this, but you could also draw U-turn ways. I only know it's there because I have been there several times myself.)

Moving one street S to osm.org/way/629622864
The W junction is clearly in the bend, so we'll merge the two nodes here.

On the E end the junction is a cross roads, not with an offset, so we'll merge these two as well and move the whole Rd S to where it should be.

The North end of osm.org/way/629622890 is parallel to a short existing stub road. In reality, these two are one and the same and this is clearly visible, if you load GPS traces and correct the imagery offset with them. Clearly, you did not do that. I have now connected the two roads.

West end of Way: 629622938: see https://www.dropbox.com/s/hoe0jxc1ys9fc90/OSMway629622890.jpg?dl=0
Road สห.2006 existed before you added the two streets on the right. You put a dog leg into the highway and the residential road crosses it before joining. Both of these are a big no-no. All you have to do is merge the two nodes and then line them up with the image. There are just about enough GPS traces on the junction N of the roundabout to let you do that. (I'm leaving this one for you to look at to learn.)

I have moved almost all roads in this area to line up with the offset.

You are drawing the roads nicely now with sweeping round curves. Once you leave out the extra nodes in the straight sections and line the whole thing up with GPS tracks it will be perfect. I see this was all done 10 days ago and it's a big improvement over edits I have seen from your colleagues a year ago.

Kind regards,
Peter.

If you don't mind, I'd like to give you some feedback on some of the mapping you have done around osm.org/way/458203539: The canal looks great.

Where you drew bridges, however, there is actually an earth dam, most likely with a culvert underneath. I have split the big canal here and in the other place a little further NW and removed the bridges.

In between the two you can see a bridge crossing the smaller drain running parallel to the roads. You haven't mapped that.

The pieces of road are unpaved and I have tagged them as such.

Thanks,
Peter.

There are several places where tracks follow stream beds and I want to be sure I have tagged these correctly. Thanks.

Correct source is:
http://www.rideasia.net/motorcycle-forum/threads/7818-23-1-16-Mae-Sariang-Two-Visit-The-Waterfall-And-Hot-Springs