Re: [Talk-us] TIGER 2010 Imports

2011-01-15 Thread Mike N

Having said that: let's start a thread here about getting the TIGER data
moving along. What steps can we take to move the shapefiles in to OSM
format? How can we collaborate on the mapping to OSM tags?


Re: TIGER 2010 tags

  To move things along, how about starting with the page 
http://wiki.openstreetmap.org/wiki/TIGER_to_OSM_Attribute_Map , to 
create a TIGER_2010_to_OSM_attribute map page?   What do we change on 
this page?  What must be added so that it is harder to assert that 
"TIGER 2010 is undocumented"?


  I'm not a tagging expert, but I have the following observations:

 1.) It seems that Nominatum (or someone?) uses tiger:zip_left and 
tiger:zip_right to create rough zip code correlations.


  2.)  Directionals - this is a bit above my head, but a good study is 
at 
http://wiki.openstreetmap.org/wiki/Proposed_features/Directional_Prefix_%26_Suffix_Indication


 so,

  fedir might map to name:prefix
  fedirs might map to name:suffix

 Even though additional names are not officially recognized in OSM, for 
any additional TIGER names:

  fedir -> name_1:prefix
  fedirs -> name_1:suffix



___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2011-01-05 Thread Mike N.
Also, it will be critical to future maintenance to keep an archive copy of the 
OSM that is deployed.  The archive can be used to create a much smaller diff 
from future versions of TIGER so that updates which have not already been 
surveyed can be applied much quicker and with minimal labor.___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2011-01-05 Thread Mike N.
>Some notes from my editing in southern California:
>
>In some places (UT comes to mind), people have begin using name:prefix to 
>(correctly) remove directional prefixes and suffixes from the name tag (see 
>>http://wiki.openstreetmap.org/wiki/Proposed_features/Directional_Prefix_%26_Suffix_Indication
> etc.).   I'm doing a little bit of this manually in CA, and will probably do 
>some larger work on it soon. If we can agree on tagging, I think it would be 
>nice to implement a complete breakout of prefix, suffix, type, etc., since the 
>info is already there in the TIGER data.

   I think the proposal was well thought out and presented.  I can't wrap my 
mind around it for the simple cases around here that just use a directional 
prefix.   But if it just a matter of grabbing the info and putting it into the 
new tag format, that sounds like a good idea.

>  or - Accept geometry within a constrained area to prevent chaos at the 
> endpoints which have been cleaned up with multiple 1-way "_link" type 
> connections / or short angle extensions, or which connect with an >interstate 
> ramp.
>
>In manually editing ramps, I consistently make the last node before the 
>motorway junction at the gore point (and then angle the ramp towards the 
>motorway by 30-60 degrees). This sometimes results in >moving the junction 
>significantly, but generally results in less complicated/conflicting 
>geometries.

   I believe that ramps will require manual fixing, especially when the new 
version contains dual carriageways replacing a single way.





>  4.)  Highlight name changes.   Review, 'accept', and the name change will be 
> applied to the existing way.
>Complications - roads with multiple names?Ability to choose 
> "name_1, _2, or _3"?
>
>I believe these tags are discouraged (by the wiki) in favor of alt_name, 
>official_name, etc. Motorway names, in particular, were a mess. When I edit 
>them, I refer to CA law and CalTrans (Cal DOT). In the >absence of evidence to 
>the contrary, these edited names should remain as they are.

  The old TIGER (and possibly new TIGER?) often had the "real name" - the one 
on the sign - alternating among multiple name_* on adjoining blocks.   I agree 
that name_1, _2, and _3 are throwaway and will not be useful in the OSM schema, 
but it is also hard to know what is real alt_name, old_name, official_name 
unless you live there or work in the highway dept.

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2011-01-05 Thread Alan Mintz

At 2011-01-05 08:00, Mike N. wrote:
...

The tool would "un-abbreviate" both the new and existing TIGER datasets so that name comparisons catch only actual changes.    In addition, the end result of the "Import, merge and review" process is that all TIGER names would be expanded and the tiger:reviewed flag would be removed.   A reference implementation is in 'expand.py' in SVN.
Some notes from my editing in southern California:
In some places (UT comes to mind), people have begin using name:prefix to (correctly) remove directional prefixes and suffixes from the name tag (see http://wiki.openstreetmap.org/wiki/Proposed_features/Directional_Prefix_%26_Suffix_Indication etc.). I'm doing a little bit of this manually in CA, and will probably do some larger work on it soon. If we can agree on tagging, I think it would be nice to implement a complete breakout of prefix, suffix, type, etc., since the info is already there in the TIGER data.
When they are still present, matching against the old imported tiger:* naming tags might end up being a better solution.

Possible steps:
  1.)   Highlight geometry changes against nearby road with same name and/or TLID.
Be aware that there were bugs in the original TIGER import that will likely make any attempt to use TLIDs futile. Additionally, roads have been combined, split, etc., sometimes yielding TLID strings that are too long. My usual response was to truncate the TLID string when it happened to me (given that I knew the information to be flawed).

  or - Accept geometry within a constrained area to prevent chaos at the endpoints which have been cleaned up with multiple 1-way "_link" type connections / or short angle extensions, or which connect with an interstate ramp.
In manually editing ramps, I consistently make the last node before the motorway junction at the gore point (and then angle the ramp towards the motorway by 30-60 degrees). This sometimes results in moving the junction significantly, but generally results in less complicated/conflicting geometries.

  4.)  Highlight name changes.   Review, 'accept', and the name change will be applied to the existing way.
    Complications - roads with multiple names?    Ability to choose "name_1, _2, or _3"?
I believe these tags are discouraged (by the wiki) in favor of alt_name, official_name, etc. Motorway names, in particular, were a mess. When I edit them, I refer to CA law and CalTrans (Cal DOT). In the absence of evidence to the contrary, these edited names should remain as they are.
The wiki previously discouraged having a name field of "State Highway n", in favor of ref="XX n", which initially makes sense when those roads have alternative names. However, I've recently been seeing roads that have no other name, are signed "State Highway n", have addresses like "q State Highway n" along them, etc., leading me to think it useful to keep the name tag when there is no other name for the road, if for no other reason than consistency with addr:street. Not sure what others have done in these cases.

--
Alan Mintz 



___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2011-01-05 Thread Mike N.
I have looked at some TIGER 2010 extracts in areas I'm familiar with, and have 
some more concrete ideas to suggest.

Ian may have time to create a customized import tool that will make the whole 
process much easier.   Here are some ideas for the workflow that should work in 
both "the boonies" where no work has been done except for US highways and 
interstates, as well as areas which have had extensive editing.

Users sign up for import on a county-by-county basis so that 2 people don't try 
to import the same region.
- Use an OSM wiki page for this, or a Google Docs spreadsheet?

Without getting too specific about the shapefile download and manipulation, "a 
miracle occurs", and the user is operating JOSM with one layer being a current 
download of the OSM region in question - a partial or complete county download, 
and another layer being the TIGER 2010 data.   

  The overall strategy is to generally apply the new geometry / names to the 
existing TIGER road data.   That retains history and any attributes that 
mappers have applied.

The tool would "un-abbreviate" both the new and existing TIGER datasets so that 
name comparisons catch only actual changes.In addition, the end result of 
the "Import, merge and review" process is that all TIGER names would be 
expanded and the tiger:reviewed flag would be removed.   A reference 
implementation is in 'expand.py' in SVN.

  The tool would operate in "batch mode" - rather than performing 2, 3, or 4 
compare operations on each way, then moving on, it is more efficient to compare 
geometry, additions, and deletions and highlight the differences.   After 
resolving geometry, next highlight name changes and resolve, etc.

Possible steps:
  1.)   Highlight geometry changes against nearby road with same name and/or 
TLID.   For each,
  Accept entire way.   New geometry would be projected onto previous way, 
including intersections and endpoints.
  or - Accept geometry within a constrained area to prevent chaos at the 
endpoints which have been cleaned up with multiple 1-way "_link" type 
connections / or short angle extensions, or which connect with an interstate 
ramp.
 Geometry could be applied where a road has been split for other reasons 
such as adding a bridge, tunnel or speed limit.   The bridge endpoints would be 
retained, and the remainder of the geometry applied.

  2.)  Highlight new roads.   Review,  'accept', and the new way would be 
automatically stitched to the nearest line at the endpoint - which should now 
be very close since the geometry has been adjusted.

  3.)  Highlight deleted roads.   Review, 'accept' and the old way would be 
automatically deleted.

  4.)  Highlight name changes.   Review, 'accept', and the name change will be 
applied to the existing way.
Complications - roads with multiple names?Ability to choose 
"name_1, _2, or _3"?

  5.)  County border line road stitching will likely be a manual operation.

  6.)  Upload and repeat until county is complete.

Complications - 
   If you're not local, you won't recognize new construction VS old / bad 
geometry; especially if TIGER matches older aerials you're reviewing against.

-
   Are there any open source implementations of this sort of tool that would 
provide more ideas?I looked at "Road Matcher" briefly, but didn't want to 
contaminate my thinking by grabbing an idea that might be patented.

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-21 Thread Toby Murray
We have discussed [1] a cheaper way of doing it to at least be a
better approximation and Antony commented on my blog saying that they
would try to incorporate this.

http://lists.openstreetmap.org/pipermail/talk-us/2010-December/004977.html
(and the following 3 messages)

Toby

On Tue, Dec 21, 2010 at 6:06 AM, Ian Dees  wrote:
> Antony from MapQuest basically said that in order to do it correctly it
> would take a full-history planet dump and they don't have the space to do
> it:
> http://lists.openstreetmap.org/pipermail/talk-us/2010-December/004976.html
>
> On Tue, Dec 21, 2010 at 12:08 AM, Alan Mintz 
> wrote:
>>
>> Anyone know if the problem with the tigerviewer map showing too much red
>> is being worked on, or where to report it?
>>
>> At 2010-12-16 06:07, Alan Mintz wrote:
>>
>> At 2010-12-15 10:04, Ian Dees wrote:
>>
>> Look at the TIGER edited map. There is *lots* of untouched TIGER data in
>> OSM:
>>
>>
>> http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B
>>
>> Do you know what the criteria are for red vs. green? Unless this came from
>> a very old dataset (<2009), I believe it's wrong. I see _way_ too much red
>> in places where I've done a lot of work.
>>
>> Here's a simple case:
>> http://open.mapquestapi.com/tigerviewer/index.html?zoom=16&lat=34.16103&lon=-117.58066&layers=B
>>
>> If you look at the blocks framed by Hermosa, Hillside, and Haven, it shows
>> only one green road.
>>
>> If you look at the area with the data view in OSM at
>> http://www.openstreetmap.org/?lat=34.15963&lon=-117.58032&zoom=16&layers=M ,
>> you'll see that I edited every street, correcting geometry, tagging source
>> and source_ref, researching naming discrepancies, adding cul-de-sacs, etc.
>>
>> In particular, if you look at
>> http://www.openstreetmap.org/browse/way/7268631/history , you see that I
>> first edited it at Wed, 27 May 2009 10:41:28 +.
>>
>> Does it maybe ignore ways where balrog-kun (or another bot) was last to
>> act, even if the way was previously edited?
>>
>> --
>> Alan Mintz 
>>
>> ___
>> Talk-us mailing list
>> Talk-us@openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/talk-us
>>
>
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>
>

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-21 Thread Ian Dees
Antony from MapQuest basically said that in order to do it correctly it
would take a full-history planet dump and they don't have the space to do
it:

http://lists.openstreetmap.org/pipermail/talk-us/2010-December/004976.html

On Tue, Dec 21, 2010 at 12:08 AM, Alan Mintz

> wrote:

> Anyone know if the problem with the tigerviewer map showing too much red is
> being worked on, or where to report it?
>
> At 2010-12-16 06:07, Alan Mintz wrote:
>
> At 2010-12-15 10:04, Ian Dees wrote:
>
> Look at the TIGER edited map. There is *lots* of untouched TIGER data in
> OSM:
>
>
> http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B
>
>
> Do you know what the criteria are for red vs. green? Unless this came from
> a very old dataset (<2009), I believe it's wrong. I see _way_ too much red
> in places where I've done a lot of work.
>
> Here's a simple case:
> http://open.mapquestapi.com/tigerviewer/index.html?zoom=16&lat=34.16103&lon=-117.58066&layers=B
>
> If you look at the blocks framed by Hermosa, Hillside, and Haven, it shows
> only one green road.
>
> If you look at the area with the data view in OSM at
> http://www.openstreetmap.org/?lat=34.15963&lon=-117.58032&zoom=16&layers=M, 
> you'll see that I edited every street, correcting geometry, tagging source
> and source_ref, researching naming discrepancies, adding cul-de-sacs, etc.
>
> In particular, if you look at
> http://www.openstreetmap.org/browse/way/7268631/history , you see that I
> first edited it at Wed, 27 May 2009 10:41:28 +.
>
> Does it maybe ignore ways where balrog-kun (or another bot) was last to
> act, even if the way was previously edited?
>
>  --
> Alan Mintz 
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>
>
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-20 Thread Alan Mintz

Anyone know if the problem with the tigerviewer map showing too much red
is being worked on, or where to report it?
At 2010-12-16 06:07, Alan Mintz wrote:
At 2010-12-15 10:04, Ian Dees
wrote:
Look at the TIGER edited map. There
is *lots* of untouched TIGER data in OSM:
http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B
Do you know what the criteria are for red vs. green? Unless this came
from a very old dataset (<2009), I believe it's wrong. I see _way_ too
much red in places where I've done a lot of work.
Here's a simple case:
http://open.mapquestapi.com/tigerviewer/index.html?zoom=16&lat=34.16103&lon=-117.58066&layers=B
If you look at the blocks framed by Hermosa, Hillside, and Haven, it
shows only one green road.
If you look at the area with the data view in OSM at
http://www.openstreetmap.org/?lat=34.15963&lon=-117.58032&zoom=16&layers=M
, you'll see that I edited every street, correcting geometry, tagging source and source_ref, researching naming discrepancies, adding cul-de-sacs, etc.
In particular, if you look at http://www.openstreetmap.org/browse/way/7268631/history , you see that I first edited it at Wed, 27 May 2009 10:41:28 +.
Does it maybe ignore ways where balrog-kun (or another bot) was last to act, even if the way was previously edited?


--
Alan Mintz 



___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Dave Hansen
On Thu, 2010-12-16 at 09:56 -0600, Toby Murray wrote:
> On Thu, Dec 16, 2010 at 7:40 AM, Mike N.  wrote:
> >I don't know how to back to git directly, plus I have no way to test
> > this, but I've attached the logic to check TIGER version as I understand it.
> 
> At first I thought this wasn't quite right but actually I guess it
> might be. The version after DaveHansenTiger will presumably always be
> 1 so there is no need to check versions there. And after balrog-kun
> gets done with them they will always be at least version 2.  I'm not
> very familiar with the work of the user Milenko. Should the version=2
> check be added there as well? 

I made the .osm files available ahead of the actual upload so that
people could review them.  Milenko took those and uploaded them by
himself, forgetting to mention to me that he'd done so.  I ended up
uploading the counties *again* over his.

-- Dave


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Dave Hansen
On Thu, 2010-12-16 at 11:08 -0500, Mike N. wrote:
>The user DaveHansenTiger posted what looked like edits /
> corrections 
> since the original upload, so the date range is best for that. 

The uploads were all done with my original DaveHansen.  When people
starting saying that they were now considering all DaveHansen uploads
part of TIGER, I decided to change.  Don't want the government getting
credit for adding picnic benches to OSM, now, do we?

So, I renamed DaveHansen to DaveHansenTiger (keeping the underlying uid
the same), and got another uid, DaveHansen.

-- Dave


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Mike N.

On Thu, Dec 16, 2010 at 7:40 AM, Mike N.  wrote:

   I don't know how to back to git directly, plus I have no way to test
this, but I've attached the logic to check TIGER version as I understand 
it.


At first I thought this wasn't quite right but actually I guess it
might be. The version after DaveHansenTiger will presumably always be
1 so there is no need to check versions there. And after balrog-kun
gets done with them they will always be at least version 2.  I'm not
very familiar with the work of the user Milenko. Should the version=2
check be added there as well?


  The user DaveHansenTiger posted what looked like edits / corrections 
since the original upload, so the date range is best for that.


 I couldn't figure out the Milenko user either to figure out if a version 
check would be the best or not.   Since they must have originally raised a 
flag, the current logic of the check must be correct.




___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Toby Murray
On Thu, Dec 16, 2010 at 7:40 AM, Mike N.  wrote:
>    I don't know how to back to git directly, plus I have no way to test
> this, but I've attached the logic to check TIGER version as I understand it.

At first I thought this wasn't quite right but actually I guess it
might be. The version after DaveHansenTiger will presumably always be
1 so there is no need to check versions there. And after balrog-kun
gets done with them they will always be at least version 2.  I'm not
very familiar with the work of the user Milenko. Should the version=2
check be added there as well?

Toby

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Toby Murray
On Thu, Dec 16, 2010 at 8:52 AM, Alan Mintz
 wrote:
>
> Looks like any road that was last edited by balrog-kun is set to red,
> regardless of where it came from originally.

Yes, yesterday we determined that the TIGER edited map only looks at
the latest version of a way. However I had assumed that the name
expansion edit only operated on TIGER ways but this seems to not be
the case. This will not affect the map on my blog since I'm explicitly
only looking at ways with a tiger:county tag but it will indeed affect
the TIGER edited map that MQ is providing. I'm not sure there is an
easy fix for this. I guess you could add maxspeed= or lanes= tags to
the ways and "take them back" :)

Toby

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Alan Mintz

At 2010-12-16 06:41, Andrew Ayre wrote:

http://open.mapquestapi.com/tigerviewer/index.html?zoom=15&lat=32.44121&lon=-110.75789&layers=B


Looks like any road that was last edited by balrog-kun is set to red, 
regardless of where it came from originally.


--
Alan Mintz 


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Andrew Ayre

Andrew Ayre wrote:

Alan Mintz wrote:

At 2010-12-15 10:04, Ian Dees wrote:
Look at the TIGER edited map. There is *lots* of untouched TIGER data 
in OSM:


http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B 
 



Do you know what the criteria are for red vs. green? Unless this came 
from a very old dataset (<2009), I believe it's wrong. I see _way_ too 
much red in places where I've done a lot of work.


There was a village that was so messed up in TIGER I ripped it all out
and redid it from my GPS traces in September 2009. I can clearly
recognize my many hours of work, and this map shows it all in red. So I
don't think it can be trusted for deciding on where to import. Sorry.

http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B 


Sorry, wrong link. Here is the right one:

http://open.mapquestapi.com/tigerviewer/index.html?zoom=15&lat=32.44121&lon=-110.75789&layers=B

--
Andy
PGP Key ID: 0xDC1B5864

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Andrew Ayre

Alan Mintz wrote:

At 2010-12-15 10:04, Ian Dees wrote:
Look at the TIGER edited map. There is *lots* of untouched TIGER data 
in OSM:


http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B 



Do you know what the criteria are for red vs. green? Unless this came 
from a very old dataset (<2009), I believe it's wrong. I see _way_ too 
much red in places where I've done a lot of work.


There was a village that was so messed up in TIGER I ripped it all out
and redid it from my GPS traces in September 2009. I can clearly
recognize my many hours of work, and this map shows it all in red. So I
don't think it can be trusted for deciding on where to import. Sorry.

http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B

Andy

--
Andy
PGP Key ID: 0xDC1B5864

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Alan Mintz

At 2010-12-15 10:22, Jeffrey Ollie wrote:
Would someone have enough resources to convert all of the TIGER 2010 
shapefiles into OSM XML data and load that into a blank OSM database with 
a new Mapnik instance (with a custom style) as well?  That way you could 
use the TIGER 2010 data as a background in JOSM or Potlatch for people to 
review the data at least and perhaps could be used to manually trace new 
data into OSM.


I'd like that. I did a quick/dirty script that takes a chunk of TIGER2009 
data and spits out KML, which I sometimes use as a reference for 
discrepancies, new road names, etc., when I don't have county or other 
official data. Quite useful in places.


--
Alan Mintz 


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Alan Mintz

At 2010-12-15 10:04, Ian Dees wrote:
Look at the TIGER edited map. There
is *lots* of untouched TIGER data in OSM:
http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B
Do you know what the criteria are for red vs. green? Unless this came
from a very old dataset (<2009), I believe it's wrong. I see _way_ too
much red in places where I've done a lot of work.
Here's a simple case:
http://open.mapquestapi.com/tigerviewer/index.html?zoom=16&lat=34.16103&lon=-117.58066&layers=B
If you look at the blocks framed by Hermosa, Hillside, and Haven, it
shows only one green road.
If you look at the area with the data view in OSM at
http://www.openstreetmap.org/?lat=34.15963&lon=-117.58032&zoom=16&layers=M
, you'll see that I edited every street, correcting geometry, tagging source and source_ref, researching naming discrepancies, adding cul-de-sacs, etc.
In particular, if you look at http://www.openstreetmap.org/browse/way/7268631/history , you see that I first edited it at Wed, 27 May 2009 10:41:28 +.
Does it maybe ignore ways where balrog-kun (or another bot) was last to act, even if the way was previously edited?

--
Alan Mintz 



___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Alan Mintz

At 2010-12-15 09:52, Mike N. wrote:

Having said that: let's start a thread here about getting the TIGER data
moving along. What steps can we take to move the shapefiles in to OSM
format? How can we collaborate on the mapping to OSM tags?


What is it you want to import from TIGER 2010 in the first place?  I'm
not convinced there's any feasible way to import TIGER 2010 while
guaranteeing that the import is more accurate than what's already
there.


 The exact thing I'm looking for in my area is

1.)   New streets - mostly from new subdivisions.
2.)  Improved geometry for road centerlines - if and where it exists in TIGER.
3.)  Updates on road names / other changes I've missed.


+1. When I spot-checked TIGER2009 centerlines for southern California, they 
were almost perfectly aligned with imagery. By comparison, the original 
TIGER (2005?) was pretty far off in most places. I've done a great deal of 
work manually, but I think there are still many places left untouched that 
could benefit from complete replacement. This would be much faster than 
manually aligning those places to imagery, even if it has to be done with 
some sort of manual matching tool.


Let's make sure we have consensus on the directional prefix/suffix issue, 
too. If I recall correctly, TIGER is now rational in this regard, and we 
should make sure we capture what's there.


Is it any better at road-type classifications, or is everything still A41? 
It seems that if an existing TIGER2005 road has been 
reviewed/edited/aligned, but is still highway=residential, and TIGER2010 
has promoted it to a higher class, we should consider doing that. I suspect 
this applies to a lot of the map - particularly the lack of tertiary roads.


--
Alan Mintz 


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Mike N.



I THINK the file you need where the rules are, is:

https://github.com/MapQuest/TIGER-Edited-map/blob/master/inc/layer-tiger.xml.inc

If you guys want to build more fine-grained rules and contribute them 
back, that would be truly awesome


I don't know how to back to git directly, plus I have no way to test 
this, but I've attached the logic to check TIGER version as I understand it.




 
   [is_tiger] = 'yes'
   &maxscale_zoom8;
   
 
 
   [is_tiger] = 'no'
   &maxscale_zoom8;
   
 
 
   [is_tiger] = 'yes'
   &maxscale_zoom9;
   
 
 
   [is_tiger] = 'no'
   &maxscale_zoom9;
   
 
 
   [is_tiger] = 'yes'
   &maxscale_zoom11;
   
 
 
   [is_tiger] = 'no'
   &maxscale_zoom11;
   
 
 
   [is_tiger] = 'yes'
   &maxscale_zoom12;
   
 
 
   [is_tiger] = 'no'
   &maxscale_zoom12;
   
 




 tiger
 
   
 (select way,
   (case when osm_uid = '7168' -- DaveHansenTiger
   and osm_timestamp::timestamp >=
'2007-09-01'::timestamp
   and osm_timestamp::timestamp <=
'2008-05-04'::timestamp
 then 'yes'
 when osm_uid = '15169' -- Milenko
   and osm_timestamp::timestamp >=
'2007-10-29'::timestamp
   and osm_timestamp::timestamp <=
'2007-12-12'::timestamp
 then 'yes'
 when osm_uid = '20587' -- balrog-kun
   and osm_version = '2'
 then 'yes'
 else 'no' end) as is_tiger
   from &prefix;_line
   where highway is not null
   order by is_tiger desc) as roads
   
   &datasource-settings;
 

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-16 Thread Mike N.

I just posted a new version of my nationwide TIGER map as well as the
logic behind the version 3 statement on my blog:


 Great analysis - that illuminates the TIGER edits nationwide.


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Toby Murray
Well I thought about this some more tonight and ran some queries on my
own database. I'm not sure how to write this in renderer-speak and it
still isn't quite perfect but if you have access to the version number
of the way in the line table that is being queried, I would suggest
adding an additional condition to those three "when" clauses that
requires the version number of the way to be under 3

I just posted a new version of my nationwide TIGER map as well as the
logic behind the version 3 statement on my blog:

http://ksmapper.blogspot.com/2010/12/updated-tiger-map.html

Toby


On Wed, Dec 15, 2010 at 5:39 PM, Antony Pegg  wrote:
> Yep, the TIGER unedited map is *not* perfect
>
> here's the style file up on Git
>
> https://github.com/MapQuest/TIGER-Edited-map
>
> the rules are in there.
>
> Basically, we can't afford to have the complete history of a way, so we have
> to go with the latest version as a guide
>
> I THINK the file you need where the rules are, is:
>
> https://github.com/MapQuest/TIGER-Edited-map/blob/master/inc/layer-tiger.xml.inc
>
> If you guys want to build more fine-grained rules and contribute them back,
> that would be truly awesome
>
> HTH
> Ant
>
>
>
> --
> Aut viam invenium aut facium
>
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Antony Pegg

Yep, the TIGER unedited map is *not* perfect

here's the style file up on Git

https://github.com/MapQuest/TIGER-Edited-map

the rules are in there.

Basically, we can't afford to have the complete history of a way, so we 
have to go with the latest version as a guide


I THINK the file you need where the rules are, is:

https://github.com/MapQuest/TIGER-Edited-map/blob/master/inc/layer-tiger.xml.inc

If you guys want to build more fine-grained rules and contribute them 
back, that would be truly awesome


HTH
Ant



--
Aut viam invenium aut facium


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Dave Hansen
On Wed, 2010-12-15 at 15:09 -0800, David Muir Sharnoff wrote:
> 
> I just tried this tool and the results it gives are incorrect.
> 
> For example, it shows this area as unedited and that is 100%
> incorrect.
> 
> http://open.mapquestapi.com/tigerviewer/index.html?zoom=17&lat=37.82347&lon=-122.19419&layers=B

Yeah, it's definitely got some issues.  Go check out Champaign County in
Illinois or Wabash in Indiana.  Those were two of the first counties
imported and are marked all green.  Some bit of the query obviously
isn't working there.  

-- Dave


___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Toby Murray
Looking at the data I think Mapquest is making the same mistake I made
on my map. All the ways in the area were last touched by balrog-kun
which was the automated edit to expand street name abbreviations. So
even though other people touched the ways before balrog-kun did, he is
considered the "owner" of the way and as such it is marked as not
having been edited. If I can figure out how to change my query to
avoid this problem then hopefully the MQ people can adapt it to their
TIGER edited map as well.

Toby


On Wed, Dec 15, 2010 at 5:09 PM, David Muir Sharnoff
 wrote:
> I just tried this tool and the results it gives are incorrect.
>
> For example, it shows this area as unedited and that is 100% incorrect.
>
> http://open.mapquestapi.com/tigerviewer/index.html?zoom=17&lat=37.82347&lon=-122.19419&layers=B
>
> -Dave
>
> On Wed, Dec 15, 2010 at 10:04 AM, Ian Dees  wrote:
>> OSM:
>> http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread David Muir Sharnoff
I just tried this tool and the results it gives are incorrect.

For example, it shows this area as unedited and that is 100% incorrect.

http://open.mapquestapi.com/tigerviewer/index.html?zoom=17&lat=37.82347&lon=-122.19419&layers=B

-Dave

On Wed, Dec 15, 2010 at 10:04 AM, Ian Dees  wrote:
> OSM:
> http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Toby Murray
On Wed, Dec 15, 2010 at 12:04 PM, Ian Dees  wrote:
>> How much of that is there, anyway?
>
> Look at the TIGER edited map. There is *lots* of untouched TIGER data in
> OSM:
> http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B


At the risk of being accused of blatant self-promotion, I would also
like to point to a blog post I wrote on the subject. It does exactly
as Mike N suggested, looking at the last editor of a way and
determining if it was automated or not. It is similar in spirit to the
TIGER edited map but aggregates the data on a county-by-county level.
http://ksmapper.blogspot.com/2010/11/nationwide-tiger-map.html

One bug I was recently made aware of on IRC is that edits that were
made after the initial import but before the "un-abbreviator" aren't
being taken into account. I will try to fix that and post a new map.

But yes, at the end of the day, there are a LOT of untouched TIGER
ways in OSM and I think a lot of them could benefit from being updated
with newer data.

Toby

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Mike N.

It would be good to document the process for loading the shape files
into a database and setting up mapnik to render from that database.  I
don't have the resources to do the whole US but I'd like to experiment
with my neighborhood.


  A quick way to take a look is to use shp-to-osm with TIGER rules, then 
load the resulting .OSM file in a separate layer in JOSM to overlay. I 
plan to take a peek, but it will be well worth waiting for the comparison 
tool scripts to be adapted and tuned to this dataset.


http://redmine.yellowbkpk.com/projects/geo/repository/show/trunk/shp-to-osm




___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Jeffrey Ollie
On Wed, Dec 15, 2010 at 12:26 PM, Ian Dees  wrote:
> On Wed, Dec 15, 2010 at 12:23 PM, Ian Dees  wrote:
>>
>> On Wed, Dec 15, 2010 at 12:22 PM, Jeffrey Ollie  wrote:
>>>
>>> I'd start with getting a wiki page set up with the types of
>>> information that is stored in the TIGER 2010 shapefiles.
>>>
>>
>> Sounds like a great idea.
>
> I created a stub page here: http://wiki.openstreetmap.org/wiki/TIGER_2010
> Beyond the metadata mappings, is there any other info that we should store
> on there?

It would be good to document the process for loading the shape files
into a database and setting up mapnik to render from that database.  I
don't have the resources to do the whole US but I'd like to experiment
with my neighborhood.

-- 
Jeff Ollie

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Ian Dees
On Wed, Dec 15, 2010 at 12:23 PM, Ian Dees  wrote:

> On Wed, Dec 15, 2010 at 12:22 PM, Jeffrey Ollie  wrote:
>
>>
>> I'd start with getting a wiki page set up with the types of
>> information that is stored in the TIGER 2010 shapefiles.
>>
>>
> Sounds like a great idea.
>
>
I created a stub page here: http://wiki.openstreetmap.org/wiki/TIGER_2010

Beyond the metadata mappings, is there any other info that we should store
on there?
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Ian Dees
On Wed, Dec 15, 2010 at 12:22 PM, Jeffrey Ollie  wrote:

> On Wed, Dec 15, 2010 at 10:01 AM, Ian Dees  wrote:
> >
> > I really don't want to run into the situation we currently have with NHD
> > where everyone is doing the conversion with different tools using
> different
> > sets of mapping files and uploading in different ways. Let's have a
> > discussion about how this should be pulled off first.
> > Having said that: let's start a thread here about getting the TIGER data
> > moving along.
>
> Would someone have enough resources to convert all of the TIGER 2010
> shapefiles into OSM XML data and load that into a blank OSM database
> with a new Mapnik instance (with a custom style) as well?  That way
> you could use the TIGER 2010 data as a background in JOSM or Potlatch
> for people to review the data at least and perhaps could be used to
> manually trace new data into OSM.
>

Yes, that's a good idea, although it would be more efficient to simply load
the shapefiles into a database and render from that.


>
> > What steps can we take to move the shapefiles in to OSM
> > format? How can we collaborate on the mapping to OSM tags?
>
> I'd start with getting a wiki page set up with the types of
> information that is stored in the TIGER 2010 shapefiles.
>
>
Sounds like a great idea.
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Jeffrey Ollie
On Wed, Dec 15, 2010 at 10:01 AM, Ian Dees  wrote:
>
> I really don't want to run into the situation we currently have with NHD
> where everyone is doing the conversion with different tools using different
> sets of mapping files and uploading in different ways. Let's have a
> discussion about how this should be pulled off first.
> Having said that: let's start a thread here about getting the TIGER data
> moving along.

Would someone have enough resources to convert all of the TIGER 2010
shapefiles into OSM XML data and load that into a blank OSM database
with a new Mapnik instance (with a custom style) as well?  That way
you could use the TIGER 2010 data as a background in JOSM or Potlatch
for people to review the data at least and perhaps could be used to
manually trace new data into OSM.

> What steps can we take to move the shapefiles in to OSM
> format? How can we collaborate on the mapping to OSM tags?

I'd start with getting a wiki page set up with the types of
information that is stored in the TIGER 2010 shapefiles.

-- 
Jeff Ollie

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Ian Dees
On Wed, Dec 15, 2010 at 12:00 PM, Anthony  wrote:

> On Wed, Dec 15, 2010 at 12:58 PM, Ian Dees  wrote:
> > On Wed, Dec 15, 2010 at 11:37 AM, Anthony  wrote:
> >>
> >> On Wed, Dec 15, 2010 at 11:01 AM, Ian Dees  wrote:
> >> > First of all, can we agree as a group to hold off on importing or
> >> > applying
> >> > any TIGER 2010 data until we come up with a way to apply changes in a
> >> > uniform and somewhat organized manner?
> >>
> >> I don't see why TIGER 2010 should be treated differently from any
> >> other imports.  If you have data that you're sure is more accurate
> >> than what's already there, and are using well-established tags, then
> >> go ahead and import.  If you're not sure if your data is more accurate
> >> than what's already there, don't import.  If you are making up your
> >> own tags, then talk about it first.
> >
> > I can't think of any US, national-level imports (other than the original
> > TIGER import, perhaps) that have gone well.
>
> That seems to me like a good argument *against* importing TIGER in a
> uniform and somewhat organized manner.
>

I disagree. The point was that previous imports have not been organized or
uniform at all which helped lead to their failure. The original TIGER import
was the closest, probably.


>
> > In the areas I've spot-checked, TIGER 2010 has better resolution and more
> > road data than untouched TIGER-imported OSM data.
>
> How much of that is there, anyway?
>

Look at the TIGER edited map. There is *lots* of untouched TIGER data in
OSM:

http://open.mapquestapi.com/tigerviewer/index.html?zoom=9&lat=40.07546&lon=-76.32&layers=B
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Anthony
On Wed, Dec 15, 2010 at 12:58 PM, Ian Dees  wrote:
> On Wed, Dec 15, 2010 at 11:37 AM, Anthony  wrote:
>>
>> On Wed, Dec 15, 2010 at 11:01 AM, Ian Dees  wrote:
>> > First of all, can we agree as a group to hold off on importing or
>> > applying
>> > any TIGER 2010 data until we come up with a way to apply changes in a
>> > uniform and somewhat organized manner?
>>
>> I don't see why TIGER 2010 should be treated differently from any
>> other imports.  If you have data that you're sure is more accurate
>> than what's already there, and are using well-established tags, then
>> go ahead and import.  If you're not sure if your data is more accurate
>> than what's already there, don't import.  If you are making up your
>> own tags, then talk about it first.
>
> I can't think of any US, national-level imports (other than the original
> TIGER import, perhaps) that have gone well.

That seems to me like a good argument *against* importing TIGER in a
uniform and somewhat organized manner.

> In the areas I've spot-checked, TIGER 2010 has better resolution and more
> road data than untouched TIGER-imported OSM data.

How much of that is there, anyway?

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Emilie Laffray
New streets can probably be detected using something like the incremental
road import for BMO in France. It should help locating those.

Emilie Laffray

On 15 December 2010 17:52, Mike N.  wrote:

> Having said that: let's start a thread here about getting the TIGER data
>>> moving along. What steps can we take to move the shapefiles in to OSM
>>> format? How can we collaborate on the mapping to OSM tags?
>>>
>>
>> What is it you want to import from TIGER 2010 in the first place?  I'm
>> not convinced there's any feasible way to import TIGER 2010 while
>> guaranteeing that the import is more accurate than what's already
>> there.
>>
>
>  The exact thing I'm looking for in my area is
>
> 1.)   New streets - mostly from new subdivisions.
> 2.)  Improved geometry for road centerlines - if and where it exists in
> TIGER.
> 3.)  Updates on road names / other changes I've missed.
>
>  I don't see this as a mass automated operation - it will be some sort of
> manual operation assisted by tools where possible.
>
>
>
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us
>
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Ian Dees
On Wed, Dec 15, 2010 at 11:37 AM, Anthony  wrote:

> On Wed, Dec 15, 2010 at 11:01 AM, Ian Dees  wrote:
> > First of all, can we agree as a group to hold off on importing or
> applying
> > any TIGER 2010 data until we come up with a way to apply changes in a
> > uniform and somewhat organized manner?
>
> I don't see why TIGER 2010 should be treated differently from any
> other imports.  If you have data that you're sure is more accurate
> than what's already there, and are using well-established tags, then
> go ahead and import.  If you're not sure if your data is more accurate
> than what's already there, don't import.  If you are making up your
> own tags, then talk about it first.
>

I can't think of any US, national-level imports (other than the original
TIGER import, perhaps) that have gone well. I know because I started or
performed several of them: county borders, NHD, etc. Local-level imports of
data from counties or states are quite a bit easier to deal with, so I'm
proposing we work on getting this national-level one closer to correct.


>
> > Having said that: let's start a thread here about getting the TIGER data
> > moving along. What steps can we take to move the shapefiles in to OSM
> > format? How can we collaborate on the mapping to OSM tags?
>
> What is it you want to import from TIGER 2010 in the first place?  I'm
> not convinced there's any feasible way to import TIGER 2010 while
> guaranteeing that the import is more accurate than what's already
> there.
>

In the areas I've spot-checked, TIGER 2010 has better resolution and more
road data than untouched TIGER-imported OSM data. I have yet to spot-check
areas that have been edited by OSM members.
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Mike N.

Having said that: let's start a thread here about getting the TIGER data
moving along. What steps can we take to move the shapefiles in to OSM
format? How can we collaborate on the mapping to OSM tags?


What is it you want to import from TIGER 2010 in the first place?  I'm
not convinced there's any feasible way to import TIGER 2010 while
guaranteeing that the import is more accurate than what's already
there.


 The exact thing I'm looking for in my area is

1.)   New streets - mostly from new subdivisions.
2.)  Improved geometry for road centerlines - if and where it exists in 
TIGER.

3.)  Updates on road names / other changes I've missed.

  I don't see this as a mass automated operation - it will be some sort of 
manual operation assisted by tools where possible.




___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Anthony
On Wed, Dec 15, 2010 at 12:08 PM, Serge Wroclawski  wrote:
> 2) Imports with existing data on the same area are nearly impossible

+10.  Unless the import is done manually, I don't foresee it being a
positive thing.  I'd love to be proven wrong, though.

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Anthony
On Wed, Dec 15, 2010 at 11:01 AM, Ian Dees  wrote:
> First of all, can we agree as a group to hold off on importing or applying
> any TIGER 2010 data until we come up with a way to apply changes in a
> uniform and somewhat organized manner?

I don't see why TIGER 2010 should be treated differently from any
other imports.  If you have data that you're sure is more accurate
than what's already there, and are using well-established tags, then
go ahead and import.  If you're not sure if your data is more accurate
than what's already there, don't import.  If you are making up your
own tags, then talk about it first.

> Having said that: let's start a thread here about getting the TIGER data
> moving along. What steps can we take to move the shapefiles in to OSM
> format? How can we collaborate on the mapping to OSM tags?

What is it you want to import from TIGER 2010 in the first place?  I'm
not convinced there's any feasible way to import TIGER 2010 while
guaranteeing that the import is more accurate than what's already
there.

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Mike N.

Imports are bad, but imagry is good.


 Good imports aren't necessarily bad - it's as if a single *very active* 
mapper covers a large area.  The future maintenance of that mapper's 
contributions is the same as if the data came from an import.



I think that's because people feel they're a part of the process when
they trace or correct tracing through imagery. We should try to create
a process that involves people as much as machines.


 I agree that some sort of trace function will be useful - especially in 
areas already mapped where the corrections may just be to the centerlines. 
This will be a good addition to the "Road Matcher" utilities already 
mentioned.



* We have a map of reviewed vs unreviewed TIGER data. We should use it.


  In many cases, I won't mark the data as reviewed if I have not surveyed 
to the ends of a road.   A test for "never edited" might just be version, 
while also disregarding known bot edits such as the un-abbreviator. 



___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] TIGER 2010 Imports

2010-12-15 Thread Serge Wroclawski
On Wed, Dec 15, 2010 at 11:01 AM, Ian Dees  wrote:
> Hi Talk-US,
> First of all, can we agree as a group to hold off on importing or applying
> any TIGER 2010 data until we come up with a way to apply changes in a
> uniform and somewhat organized manner?

Preach it brother!

> I really don't want to run into the situation we currently have with NHD
> where everyone is doing the conversion with different tools using different
> sets of mapping files and uploading in different ways. Let's have a
> discussion about how this should be pulled off first.

> Having said that: let's start a thread here about getting the TIGER data
> moving along. What steps can we take to move the shapefiles in to OSM
> format? How can we collaborate on the mapping to OSM tags?

I think we've learned two things from the last few years of imports:

1) Imports are hard to do right

2) Imports with existing data on the same area are nearly impossible

3) It only takes a few minutes to import, and years (or longer) to
clean up a bad import

4) Imports' impact on the community is controversial.

I know I've talked with you personally about this problem and we've
floated a few idea, but here are some thoughts:

* Bing Imagry has been Good

Imports are bad, but imagry is good.

I think that's because people feel they're a part of the process when
they trace or correct tracing through imagery. We should try to create
a process that involves people as much as machines.

* We have a map of reviewed vs unreviewed TIGER data. We should use it.

* Other examples of import comparison tools exist

Let's try to look at other tools like Musical Chairs or the Canadian
Import tool and if not use them directly, see what we can learn.

* We have our own dataset from GPX data

We often forget we have our own datasets in the form of GPX traces.
Let's not forget that.

There's nothing direct or practical here.

What do people think of a phone or IRC meeting, getting everyone on at
the same time?

- Serge

___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


[Talk-us] TIGER 2010 Imports

2010-12-15 Thread Ian Dees
Hi Talk-US,

First of all, can we agree as a group to hold off on importing or applying
any TIGER 2010 data until we come up with a way to apply changes in a
uniform and somewhat organized manner?

I really don't want to run into the situation we currently have with NHD
where everyone is doing the conversion with different tools using different
sets of mapping files and uploading in different ways. Let's have a
discussion about how this should be pulled off first.

Having said that: let's start a thread here about getting the TIGER data
moving along. What steps can we take to move the shapefiles in to OSM
format? How can we collaborate on the mapping to OSM tags?

-Ian
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us