[OSM-talk] fixme=name

stevea steveaOSM at softworkers.com
Fri Mar 13 01:48:17 UTC 2020


There is a lot going on in this topic:  primarily, a fairly large, potentially unknowably large semantic of meanings the author of a fixme tag meant when created.  Let's be careful as we interpret these.

Assumptions by the recipient of that tag should be cautious, lest they wrongly predict the intent of the author.  I find fixme most useful when it paints a bright line forward of what needs fixing in a map datum (a node with tags, for example).  If the fixme tag is poorly written, ambiguous or the like, please, (without complaint, please) do your best to interpret how it was meant to be helpful and improve the tags (location, whatever) of the datum.

Data entered into our map (especially by beginners or maybe as part of a quickly-assembled HOT group, for example, which might include novice mappers) are not always perfect.  They should be at the very least "good," hopefully very good or excellent and during some happy fraction of data entry might even be called "perfect."  Yes, some OSM data remain at Version 1 of their History, they were that good at entry, "no improvements required" since.

Data in our map change, they live.  The lifecycle sometimes includes "fixme" tags.  Such tags, such a part of being in the data lifecycle means we must learn to do our best at improving these as we see them.  Discussion like this helps, but I feel sure "these should have entered perfectly to begin with" is a downward spiral I don't want to walk.  Let's improve data, even as we talk about how we improve data while we realize people can "mean" a lot of things with sloppy (yes, correct, if a bit harsh) data entry.  Cut this sloppiness a bit of slack, do your best to understand what they meant by that and improve what you can.  Then, our map continues to grow better (in both data and data quality).  Yes, this is a version of "perfection is the enemy of the good."

I know I walk right up to a line here of "don't put junk into our map, or stuff that'll get crufty over time" and yes, that is a real concern, I realize.

SteveA

> On Mar 12, 2020, at 4:29 PM, John Whelan <jwhelan0112 at gmail.com> wrote:
> 
> But then you're often talking about a HOT mapper who might not have done any mapping for three years.
> 
> I must confess when validating HOT projects if I saw a mapper had done something glaringly wrong and it was more than a week before noting the response rate to a changeset message was less than 1% I may have simply corrected the work.  
> 
> I haven't been on the ground in Africa but I suspect the number of villages a kilometre apart connected by a one kilometre motorway is not high.  If have I reclassified someone's private motorway  I apologise.
> 
> I think we have talked around the subject enough and it should be left to the local mappers to decide what to do.
> 
> Thanks John
> 
> Mateusz Konieczny via talk wrote on 2020-03-12 7:17 PM:
>> 
>> 
>> 
>> Mar 12, 2020, 23:54 by dieterdreist at gmail.com:
>> 
>> 
>> sent from a phone
>> On 12. Mar 2020, at 17:42, Marc M. <marc_marc_irc at hotmail.com> wrote:
>> 
>> we may delete all fixme=name for all object without a name tag
>> 
>> 
>> I agree that fixme=name for missing names is pointless and could be removed (although Andy has a point about not knowing someone’s workflow, so for safety you could spare those added in the past 1-3 months). If there are other questions concerning the name it would obviously have been better if they had been more explicit about these potential problems in the fixme tag, but still, if someone has added a fixme=name on an object with a name we should keep it.
>> Though unspecific fixme=name is not very useful, I would ask in changeset that added this
>> what exactly is wrong/suspicious if I would encounter it during editing.
>> 
>> And yes, fixme=name on object without name is likely completely useless.
>> But I would also consider asking people adding it whatever it is used in any way.
>> 
>> 
>> _______________________________________________
>> talk mailing list
>> 
>> talk at openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
> 
> -- 
> Sent from Postbox
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk




More information about the talk mailing list