[OSM-talk] Proposed new status for tags in the wiki: "import" for undiscussed tags that were only used by an import

[stevea]

See, data always have a backstory.  Thinking you know what it is, or that you can improve upon OSM by erasing existing data that has a backstory, hmmm, give that one a good, long think first before you do anything.  Discuss with others, research, think about past, present and even future data/tagging schemes that might truly improve what you attempt to improve.  Doing this is complex and deserves complex treatment, not a gloss-over and quick action.

SteveA

> On Mar 17, 2020, at 4:38 PM, stevea <steveaOSM at softworkers.com> wrote:
> 
> I would like to stress once again how easily it is for intended semantics of what is meant to be tagged, "improve-tagged" or "tag-modernized so that people understand the historical context of this tag" to diverge from the semantics that OTHER volunteers / contributors to OSM glean from these.  It is SO easy for these to be far apart and people to misunderstand one another.
> 
> This entire endeavor is fraught with peril and is one of the most slippery and dangerous (in the sense of hurt feelings due to misunderstandings, usually unintentional) in any scheme that has to do with "tagging," as in OSM with our tags (and their meant-to-be-static, though actually change through time) semantics.
> 
> Please, let's better understand the very wide aspects of what's going on here:  people invent a tag to mean "something" and perhaps it does for a while, but it might get stretched with time and might morph to something else.  And/or other tags emerge that better or "more newly" describe a scheme to tag.  Meantime, there are rendering issues (some positive, some negative) happening in parallel.  Even as people are mostly well-intentioned, this process (especially as the project gets more mature and stretches across generations of this happening, each cycle might be years or a decade) really is complex and gives rise to all kinds of tangly, snarly misunderstandings.  Tread lightly, be cautious, try to be open-minded, have both a historical understanding of how "meanings change over time" (even as we wish they didn't" and "renderers change over time" (not always exactly in-line with tagging schemes) as well as a willingness to expand context to the future.
> 
> And perhaps several other things I'm forgetting to mention... and we MIGHT be able to better solve these issues.  We can solve them, we have to be smart, patient and knowledgable about our past, looking to the future and aware of how things drift and evolve.  That's tough, but doable.
> 
> Whew!
> SteveA
> 
>> On Mar 17, 2020, at 4:17 PM, Joseph Eisenberg <joseph.eisenberg at gmail.com> wrote:
>> 
>> Unlike some of those who responded, I was not intending this status to
>> be a "mark of shame", but rather informative.
>> 
>> As mentioned, some imported tags like "gnis:feature_id=*" are useful
>> to keep the Openstreetmap database object directly linked to an object
>> in an external database.
>> 
>> That's why I am not suggesting the use of "deprecated" or "obsolete",
>> since these tags should not necessarily be removed.
>> 
>> The main reason to mark them is so that mappers and database users
>> will understand where the tag came from, and it may suggest that
>> mappers will not want to add these tags to objects in the future,
>> unless they are also importing features from the same source.
>> 
>> Besides the tags mentioned above, I was thinking about tags like
>> "object:postcode=" and "object:housenumber" - this tag is only used in
>> Germany on "highway=street_lamp" features which appear to have been
>> imported mostly in 2015: http://wiki.openstreetmap.org/wiki/Key:object
>> and see taghistory:
>> https://taghistory.raifer.tech/#***/object:postcode/ and
>> 
>> So, though the usage numbers are moderately high, it is helpful to
>> know that these tags are not really being used, except in that
>> particular context. Apparently it makes sense in the context of the
>> addressing system there, at least according to the mappers who
>> imported the objects.
>> 
>> If a tag which was first used in an imported then becomes popular and
>> used frequently by. mappers for new or updated features, then it could
>> change to "in use" or even "de facto", just like a "draft" or
>> "proposed" tag can change status due to usage over time.
>> 
>> So, just like the status "draft", the status "import" would be a hint
>> for mappers and database users, but would not suggest that the tag
>> needs to be removed, and it might change status in the future based on
>> use by mappers.
>> 
>> -- Joseph Eisenberg
>> 
>> On 3/18/20, Jmapb <jmapb at gmx.com> wrote:
>>> On 3/17/2020 10:52 AM, Wayne Emerson, Jr. via talk wrote:
>>>> However, among your examples you cite "gnis:feature_id=*" The wiki
>>>> page for this key notes:
>>>> "Unlike other imported tags such as gnis:created=* and
>>>> gnis:import_uuid=*, gnis:feature_id=* is meaningful beyond the import.
>>>> In fact, some mappers actively add gnis:feature_id=* to features to
>>>> cite a verifiable source for the POI's existence or its name."
>>> 
>>> Agree with clemency for gnis:feature_id -- it's handy to be able to
>>> crossreference features with the GNIS database, which you can search by
>>> feature id here: https://geonames.usgs.gov/apex/f?p=138:1:0:::::
>>> 
>>> J
>>> 
>>> 
>>> _______________________________________________
>>> talk mailing list
>>> talk at openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk
>>> 
>> 
>> _______________________________________________
>> talk mailing list
>> talk at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
> 
> 
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk