[OSM-dev] Fwd: Name suggestions for iD

[Aaron Lidman]

Hi Serge,

Yeah we noticed the tagging issue but there are very few cases of it so 
we're really not too concerned about it. It just resolves to a singular 
'correct' name no matter the tagging.

John and I briefly talked about how this could be made into a QA tool 
outside of iD with the goal of fixing names and tagging in existing 
data. Basically there are some weird names in the topNames.json list 
that seem really specific, like maybe they were only used a single user 
(or maybe a few), but we don't have any data to see how widespread the 
problem is. If we also keep track of users who created these names, 
along with the node/way id we could easily track down issues and then 
determine if they should be corrected or if the 'incorrect' usage is 
actually legitimate. That's down the line somewhere, we're busy with 
other stuff, but the idea is to solve that problem upstream in the data 
instead. That would require it's own project and specific discussion 
about particular use cases like Dunkin Donuts, and more discussion 
before going ahead and changing any data. Thoughts?

On 11/1/13 4:29 AM, Serge Wroclawski wrote:
> Forgot to his reply all.
>
> - Serge
>
> ---------- Forwarded message ----------
> From: Serge Wroclawski <emacsen at gmail.com>
> Date: Fri, Nov 1, 2013 at 7:29 AM
> Subject: Re: [OSM-dev] Name suggestions for iD
> To: Aaron Lidman <aaronlidman at gmail.com>
>
>
> Aaron,
>
> Thank you for taking this on. I'd done a lot of work on this but with
> the hope that we could make it work across editors. That turned out to
> be impossible.
>
> The other part, of course, is that names are only half of the issue,
> the other half is tagging. Is Dunkin Donuts a coffee shop, donut shop,
> cafe, etc.
>
> I'm wondering if adding those tags and then making a simple "Quick
> add" might be the right way to handle this, or if that's too large in
> scope.
>
> - Serge
>
> On Thu, Oct 31, 2013 at 3:43 PM, Aaron Lidman <aaronlidman at gmail.com> wrote:
>> Hello everyone,
>>
>> I've been working on a project to improve naming features in iD by auto
>> completing common names when a user is typing them in. The goal is to
>> suggest the most correct canonical version of a name for a certain type of
>> feature rather than having the many various version of 'McDonalds',
>> 'McDonald's', 'MacDonalds' that we currently have. This is not about editing
>> existing OSM data but instead aiding the user in entering what they mean in
>> the most correct way that is already being used in OSM.
>>
>> Some previous discussion on this topic:
>> https://github.com/systemed/iD/issues/1752
>> https://github.com/systemed/iD/pull/1911
>>
>> The project is here: https://github.com/osmlab/name-suggestion-index
>>
>> We want to get some community involvement in helping to determine what are
>> the most correct names by going through the list of top names currently in
>> OSM (topNames.json) and the any names that might be less correct derivations
>> of that name that should not be used. See the README for more details and
>> examples. Eventually iD will use this repo as a source for name suggestions,
>> just like the
>> [editory-imagery-index](https://github.com/osmlab/editor-imagery-index) is
>> used for available imagery.
>>
>> Any thoughts and especially contributions appreciated.
>>
>>
>> _______________________________________________
>> dev mailing list
>> dev at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/dev