[OSM-talk-be] generalized survey and consequences

[André Pirard]

On 2014-06-09 19:14, Marc Gemis wrote :
> I don't understand the real value of survey:date or whatever other key
> you want to use to indicate the last survey date. Sometimes things do
> not change for many years, sometimes they change the day after the
> survey was done.
The idea is simple.

Assuming no change for many years, different persons could come say
every week to check for a change.
The survey key tells the next one to come that it's not really necessary
to check again.
Or, put more properly, serves to establish a priority list so that long
dated surveys are refreshed first.

One cannot help a change happening just after a survey, it will wait its
turn.
> I'm struggling with this "resurvey" item as well. I've been mapping my
> area for 3 years now. Many things have changed: new walking paths, new
> streets, new one ways (it changed 3 times in the same street now).
> Sometimes I remember that things have changed, sometimes not (e.g. new
> benches, a shop moved next door, a bakery that changed name). And that
> is for an area that I know rather well.
It would be a valid point to say that some features need more frequent
survey.
But it's up to the contributors to select them in their queries to apply
a priority.
It's complicated enough already to have it worked simply.
> For places where I come less regularly I will probably miss all those
> changes, unless I do a complete resurvey of the area.
>
> What if the previous mapper doesn't survey the same things I map ?
> What's the use of the survey date in that case?
That is the point I raised that unlike source=... survey:date applies to
all aspect of a map element.
> For me, in most cases the source is useless. If I notice something
> that is different now, it doesn't matter what the previous mapper
> wrote as source. I have to change it.
>
> I do tag my changesets with survey:date for quite a while now,  but
> that is to find back my notes. And this tag is mentioned on the main
> map feature
> page http://wiki.openstreetmap.org/wiki/Map_Features#Annotation . 
>
> Furthermore, as you indicate the survey:date is on a changeset, but
> which area was revisited, what did the mapper survey ? You don't know.
>
> So in summary: for me all source tags are pretty useless, and I need
> to resurvey any area completely each time I visit it.
>
> regards
>
> m
>
>
> On Mon, Jun 9, 2014 at 6:01 PM, André Pirard <A.Pirard.Papou at gmail.com
> <mailto:A.Pirard.Papou at gmail.com>> wrote:
>
>     On 2014-06-09 11:59, Glenn Plas wrote :
>>     On 09-06-14 08:31, André Pirard wrote:
>>>     Hi,
>>>
>>>     Some data of the map changes often, in particular what's on the
>>>     road: traffic signs, bus lines etc.
>>>     It would be interesting if someone tackling a region could
>>>     determine what in his interests was checked the longest ago. 
>>>     Hence the need for a date at the beginning of the data that is
>>>     not of the source of information if any but that indicates when
>>>     that source, visual observation or other was still current last.
>>>     The someone would deal with the oldest in priority and update
>>>     that date if that can be said. The data field of the query
>>>     result would be sorted to determine the oldest ones.
>>>     Is the source:survey date appropriate for that, pardon my
>>>     limited English ...
>>     Hi Andre,
>>
>>     I think that the correct key is survey:date.
>     Thank you for replying and confirming that high precision is
>     needed in this too fuzzy OSM world.
>     I found no "survey:" key, if I look for
>     http://wiki.openstreetmap.org/wiki/survey, it falls back on
>     key:source  http://wiki.openstreetmap.org/wiki/Survey#Annotation
>     (which is a non existing label).
>     What I'm talking about is Key:source
>     <http://wiki.openstreetmap.org/wiki/Key:source> and more
>     specifically its phrase "source:name
>     <http://wiki.openstreetmap.org/wiki/Key:source:name>=survey 10
>     November 2012".
>     That is, data consisting of lowercase "survey" followed by a
>     mandatory one and only single blank...
>
>>     KeyMapper 3 also uses it.  
>     Using an URL to spare 50 people a search
>     <http://wiki.openstreetmap.org/wiki/Keypad-Mapper_3#new_features_version_3.1>,
>     indeed, thanks.
>     So, either we warn Keymapper that they use unofficial tagging that
>     escapes an overpass search,
>     or I still have to learn what many people are trying to teach me:
>     that OSM is nothing but fuzzy (sending cars the wrong one-way) and
>     that the overpass query has to be extra huge.
>     survey:date is not providing for telling what has been
>>     It's a good idea to start including this in my regular edits, I'm
>>     going to add those as well.  There is added value in it.  I think
>>     it's best to do this on the changeset but that might go unnoticed
>>     when editing, also in josm.
>     It's useful in JOSM to save ourselves checking the same element 36
>     times but mostly with overpass to make oneself a to do list.
>
>>     But that tag on every object seems like overkill.
>     Of course, only what often "changes without notice".
>
>     At first sight, the overpass API is able to use a regexp to look
>     for data but not for keys.
>     Any trick?
>
>     André.
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-be/attachments/20140609/fa759085/attachment.htm>