OpenStreetMap

Copying from OS Locator

Posted by Vclaw on 16 August 2011 in English (English)

It is not helpful to copy streetnames from OS Locator for places you have never visited.
It is not helpful to change correct streetnames, which have been properly surveyed.
It is not helpful to copy all of the blatant errors from OS Locator into Openstreetmap.

It may be increasing the numbers for 'completeness', but it is not improving the map. It makes it much harder to find actual errors, or places that need surveying.
Plus it is discouraging people from actually going out and surveying their area.

Comment from EdLoach on 16 August 2011 at 06:37

I would say that if done carefully then having a source:name=OS_Locator layer is no worse than having a nonames layer. Omitting such source tags would certainly be "not helpful", but if done carefully the map can be improved to a stage where the OS Locator ITO overlay shows roads where names differ between Locator/Streetview, or differ between Locator and what is already there (I wouldn't change an existing name without a survey). I have done a bit of armchair mapping, but set a few ground rules before doing so, documented on my user wiki page here:

http://wiki.openstreetmap.org/wiki/User:EdLoach#Armchair_mapping

It is a lot of effort though (and in many cases I was naming roads that were traced from NPE, as well as adding a few newer ones), and I've currently moved onto fixing the existing administrative boundary relations in the British Isles (not adding new ones, just fixing those that have already been started but not finished).

Hide this comment

Comment from EdLoach on 16 August 2011 at 06:37

Or even source:name=OS_OpenData_Locator, sorry...

Hide this comment

Comment from Hawkeye on 16 August 2011 at 08:06

I have been adding lots of names from Locator in Scotland but always added source:name and tried my best not to disturb what's already there. adding not:name when it has been surveyed and FIXME if I think it's worth double checking. I don't think it necessary puts people off contributing and makes the map much more usable. I take on board what you are saying and will continue with caution.

Hide this comment

Comment from Central America on 16 August 2011 at 08:46

If we wait till there is someone that can visit each location, the map will not be completed in the next 20 years. If we have a relatively complete data set where we are sending information back to OS to correct there data then that is benificial. If we have a relatively complete data set then we can use that to lobby councils and government for their data sets with the "official list of road names'. I know that there are more mistakes with the Gaelic names. We will be able to correct them even if historically some places in scotland have 15 individual names for the same place. There is much more to the map than road names, the road map is only the skeleton for the rest of the data.

Hide this comment

Comment from Harry Wood on 16 August 2011 at 09:51

I've just added something to this section.

The debate armchair mapping having a discouraging effect upon real mapping, is an old one which keeps playing out for different data sources. We've discussed in relation to aerial imagery. I remember debates about using Yahoo in London in the early days. And we've discussed in relation to imports of course. We've even pondered whether the Milton Keynes mapping party was the right thing to do because new users and real mappers might be discouraged by all the real mapping we'd done! It's a tricky one. Of course the decision gets taken away from us when just a few people blat in massive area, perhaps without realising they might be doing something bad. It'll probably happen that way with OS Locator sooner or later. I suggest we decide what all the advice should be on that wiki section there.

Hide this comment

Comment from Vclaw on 16 August 2011 at 16:33

"20 years to complete the map" is nonsense. If there aren't many contributors in a particular area, maybe you could spread the word about OSM, and encourage others to contribute. Or you could go and visit those places yourself. And then you will get much more detail than you can by tracing.

If you want a "relatively complete" (low quality) data set, you can just use OS Opendata. There's not much point in dumping all of that into OSM (introducing more errors in the process).
Its much more useful if OSM is a different data set, then they can be compared to find possible errors.

Hawkeye: I don't think adding not:name tags without surveying/checking them is useful. You don't know whether its a mistake in the original survey, or if its a valid alternative name for that street. I think its best to have these highlighted by comparison tools, as places that need checked.

Hide this comment

Comment from robert on 17 August 2011 at 00:00

Contentious issue.

Some of us actually want to, you know, _use_ the map. And sourcing street names from Locator is a way of quickly getting areas actually usable for something.

Hide this comment

Comment from chillly on 17 August 2011 at 11:32

As you say, contentious. Some of us want the very best map possible, just copying other people's stuff, errors and all, doesn't give us that. Adding OS:locator names to country roads, when there is no sign is not good. When the Sat-Nav says 'turn left into Bluestone Bottoms' but the road has no sign it can be confusing. I think aerial imagery, imports and ooc maps help a lot with alignment, give us access to awkward places and boundaries but do not replace the huge value of a survey on the ground, which is the best way to collect names imho. OS Locator does provide a great way to find things that have not been surveyed - when a new version comes out OS have helpfully added new stuff that ITO World and Musical Chairs then highlight so we can then visit and survey. Chasing the completeness lists might be fun, but it shouldn't be at the expense of accuracy or quality. Lint tools are not pointing out errors, just anomalies which are usually best resolved by a visit. "20 years to complete the map"? It will never be complete, but it will be very, very useful in much, much less time than that.

Hide this comment

Leave a comment

Parsed with Markdown

  • Headings

    # Heading
    ## Subheading

  • Unordered list

    * First item
    * Second item

  • Ordered list

    1. First item
    2. Second item

  • Link

    [Text](URL)
  • Image

    ![Alt text](URL)

Login to leave a comment