[Tagging] Tag approval process or its absence (was: Voting for Relation type=waterway)

LM_1 flukas.robot+osm at gmail.com
Mon Feb 20 23:31:41 GMT 2012


2012/2/20 Peter Wendorff <wendorff at uni-paderborn.de>:
> Hi.
> Partly that's true: non-standard tags are kind of the hell for coders, who
> have to implement lots of variants sometimes for similar information
> retrieval.
> But I vote for doing that
> 1) only manually in the database, step by step.
> or
> 2) only to add redundant tags automatically, not to remove the "deprecated"
> tags. Even that has to be done carefully and after a discussion, where the
> majority of the discussing people agree to the new variant.
I agree

>
> There are very very very few tags, that have the same meaning. Often tag
> meanings intersect, and two variants can express similar or even the same
> issues, but on the other hand most often different meanings can occur in
> combination with third tags.
> If it is possible to create a bot that "fixes" this, why not provide the
> part of a osmosis command chain to fix it in the preprocessing for
> applications?
If we can have "clean" and easy to use data compared to the need for
complex pre-processing every time it should be used, I vote for clean.
>
> Yes: we have to get rid of "old" or "deprecated" tags, but OSM does not know
> about "deprecation" - deprecated is, what's not in use anymore.
> But as often: Use is the amount of work and the number of people knowing and
> using a tag, not the first bot using or deleting a special tag.
>
> Therefore:
> 1) try to use the tags as they are, even with software
> 2) where it get's too cluttered, add your preferred tags redundant to the
> objects, keeping the other tags - and discuss this with the community before
> 3) speak with users, who add the "old" variant to convince them to use the
> new variant instead
> 4) speak to other software developers to support the new (and for editor
> software "deprecate" the old) variant.
> 5) Document it in the wiki: especially mention it in pages referring to the
> old scheme.
Again I agree

>
> Overall: be careful at forcing a new tagging scheme, where a different one
> is in favour of the majority, and don't change documentation, as long as
> that's the case.
>
> regards
> Peter
>
> Am 20.02.2012 22:59, schrieb LM_1:
>
>> 2012/2/20 Chris Hill<osm at raggedred.net>:
>>
>>> Flattening the tag structure by homogenising tags is destroying the fine
>>> detail, sometimes carefully crafted by mappers and I will continue to
>>> speak
>>> out against mass edits that attempt to do just that.
>>
>> I have to disagree. If the tag structure is not homogenised, it makes
>> the data useless. Non-standard and/or undocumented tags are impossible
>> to process in any reasonable way, even if they look perfectly complete
>> and informative to human.
>> The possibility of free tags is great, but once some tagging style
>> proves as usable (and better than any other), it should become a
>> standard and used exclusively (or be challenged by a better one
>> later).
>> Lukáš Matějka (LM_1)
>>
>>
>> 2012/2/20 Chris Hill<osm at raggedred.net>:
>>>
>>> On 19/02/12 23:38, Steve Bennett wrote:
>>>>
>>>> On Mon, Feb 20, 2012 at 2:53 AM, Chris Hill<osm at raggedred.net>    wrote:
>>>>>
>>>>> I do not agree with the whole basis of this thread.
>>>>>
>>>>> There are no such things as approved tags, tagging is open and people
>>>>> are
>>>>> free to use *any* tags they like.
>>>>
>>>> ...
>>>>>
>>>>> Advertise your ideas and encourage acceptance. Show how well it works
>>>>> any
>>>>
>>>> How would you know whether a tag had "acceptance"? Wouldn't
>>>> documenting it somewhere make sense? Maybe...in a wiki?
>>>
>>> I did say document and discuss the OP.
>>>
>>>> What would you
>>>> call "acceptance"? Would "approved" be a reasonable synonym for that?
>>>
>>> No. It implies some official status that leads people to remove other
>>> tags,
>>> sometimes with mass edits.
>>>
>>>> The wiki and (currently broken) approval mechanism is not some
>>>> horrible bureaucracy that exists to ruin your life. It's there so we,
>>>> as a community, can document the tags we use, and agree on how we use
>>>> them. While it's ok to spontaneously invent a new tag and use it to
>>>> solve your current problem, you can surely see the benefits of
>>>> everyone eventually converging on the same tag?
>>>>
>>>> And if so, what would you do with all the old tags that people used
>>>> before you converged? Wouldn't you "deprecate" them?
>>>
>>> No, some tags will wither away, fine. Some seemingly similar tags will
>>> exist
>>> side-by-side and that is fine too. Most importantly, distinctive
>>> differences
>>> can emerge too.
>>>
>>> Just think this through. Approval implies some sort of enforcement,
>>> without
>>> enforcement what is the point of approval? Just who would make this
>>> enforcement happen and how? What would that do to an open project? If
>>> only
>>> approved tags are used then how would mappers map what they actually see?
>>> Wait weeks for some committee to discuss, argue and approve or reject the
>>> tag? If you are free to use any tag, what is an approval process for?
>>>
>>> If approval or 'acceptance' means a tag is rendered or used in a router
>>> or
>>> whatever then which tool do you mean? There are hundreds run by OSM and
>>> other organisations, companies and individuals.
>>>
>>> Flattening the tag structure by homogenising tags is destroying the fine
>>> detail, sometimes carefully crafted by mappers and I will continue to
>>> speak
>>> out against mass edits that attempt to do just that.
>>>
>>>
>>> --
>>> Cheers, Chris
>>> user: chillly
>>>
>>>
>>> _______________________________________________
>>> Tagging mailing list
>>> Tagging at openstreetmap.org
>>> http://lists.openstreetmap.org/listinfo/tagging
>>
>> _______________________________________________
>> Tagging mailing list
>> Tagging at openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/tagging
>
>
>
> _______________________________________________
> Tagging mailing list
> Tagging at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/tagging



More information about the Tagging mailing list