[OSM-talk] Prolification of the amenity tag

[Andy Robinson]

bvh wrote:
>Sent: 29 November 2006 7:25 PM
>To: talk at openstreetmap.org
>Subject: Re: [OSM-talk] Prolification of the amenity tag
>
>On Wed, Nov 29, 2006 at 04:25:51PM +0000, 80n wrote:
>> The current ratio of maps to renderers is about 0.0000001:3 ;-)  We have
>> about 0.0000001% of the planet mapped.  There are three renderers
>available
>> on the OSM front page (Mapnik, Osmarender, Mapnik WMS-C), and I am sure
>> there are others.
>>
>> What is unrealistic about a ratio of say 1:1000?
>
>I am not following you there? Maps to renderes? We are barely able
>to render 0.00001% of the planet 3 times. So we are certainly not
>able to render 100% of the planet a multitude of times like you
>seem to propose?
>
>> Most mashups of geo data that I have seen comprise a pre-rendered base
>map
>> (usually Google maps or similar) with another dataset overlayed on top.
>I
>> think this is what most people imagine when talking about mashups.
>
>Yes.
>
>> The opportunity to access the source geodata in OSM enables multiple
>> datasets to be merged before rendering.  Whether the data comes from OSM
>or
>> somewhere else is unimportant - the ability to merge it before rendering
>> *is* important.  A very easy way to achieve that is to store both
>datasets
>> in OSM, using tag prefixes, if neccessary, to keep them separate and
>> managable.
>
>Why is this important? What extra possibilities does this give?
>And isn't it more important to be able to
>do a query like "give me all the banks that have a foreign-cash
>exchange service in Rotterdam", which would be more convenient to
>implement if the datasets are kept seperate.
>

If you break that query down there are locative references and there are
information references. The locative references should definitely be part of
OSM. That's what OSM is for, ie gathering. storing and delivering
geolocation data. So the location of the Banks in Rotterdam in this
instance. Provided you can match the referencing of this OSM locative data
with the corresponding extended information data then there is no
requirement to store the latter in OSM, however it's a lot easier to manage
a single data source with all of these attributes applying to the same
entity than separate ones and so I can't really see why the extended data
cannot be part of OSM.

If we look at other mapping data organisations, like the OS for instance,
the amount of information data within the Mastermap must be huge, its not
all locative data but it is data linked to locative data. Why would OSM need
to be any different?

There is absolutely no reason why someone cannot have a separate database
that is based in part or in whole on the OSM data set representing banks or
any other subset of the full data.

Cheers

Andy


>cu bart
>
>_______________________________________________
>talk mailing list
>talk at openstreetmap.org
>http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk