I too "found the light" after following the Talk mailing list, but had
already done ~8 months of tracing :-( Slowly fixing that up now.
I too think that data that has no source/source:name/source:ref tags
etc. is not trustworthy, but is better than nothing (because someone
can now see that it needs to be verified, & goes out and does a survey).
Your example is a bit tough, I don't think there is consensus yet on how
to mark up the extra details. Given that, I'd expect:
source=yahoo name=blah source:name=MMBW
then something like:
oneway:source=knowledge traffic_calming:source=gps
Smushing all sources into the one source=* tag is probably the worst way
of doing it
PS. It seems it's an anomaly that source:name isn't name:source, I think
it was for grouping source tags when sorted alphabetically. Looking
at Tagwatch, both are popular, but source:name is more so. I believe
some tools handle either format.
IIRK, one of the OSM founders said on Talk, source=gps is better than
source=survey, as survey implies a professional survey with surveying
equipment, for example from public domain government data.
BTW, Potlatch has source=* shortcut keys, JOSM & Merkaartor probably do
too.
I don't think it's intractable problem, with crowdsourcing, good tools &
leadership on "best practices" by the experienced, it's solvable.
I think of source:name as a series of levels, where you can overwrite a
source with a higher one:
1. source:name=image (GPS tagged photo's the best)
2. source:name=voice
3. source:name=survey (poorly named, but for a notepad etc.)
4. source:name=knowledge
5. source:name=MMBW/Collins1936 (Out-of-copyright)
6. source:name=historical
I use survey, writing in my PDA's notepad application & save as PNG's on
my computer as an evidence trail. I should probably use
source:name=notepad to be explicit. When you think about it, almost all
could be doctored (from a copyright source) except for photo's, so I
think that's the ultimate goal (all street signs/POI have a
corresponding GPS tagged photo of them), yep, a lot of work.
Maybe I should get around to buying a GPS :-)
source=* is different, since just because you tweak a few road bends from
a GPS track doesn't mean it wasn't derived from Yahoo tracing.
I'd suggest source=yahoo;gps in that case.
Of course I'm not a copyright lawyer, but that seems to make sense to me.
I can't imagine a scenario where is would be ok to change the source on
a road from a government import, just because someone drove down the
road wih their GPS. It still is derived from the Government data.
Cheers,
BlueMM
--- On Sun, 18/1/09, Matt White <mattwh...@iinet.com.au> wrote:
> I'm also a serial offender for not using the survey tag, but again, I
> upload all my traces.
>
> I do a bit of yahoo tracing as well, and don't directly mark that
> either...
>
> But then again, I don't go and screw around with any existing nodes and
> ways unless I've been there... however I have been known to add the
> road way in off a random trace that wasn't mine if I've been in the
> area (sans name generally) and the trace quality is good enough.
>
> It's an intractable problem - forcing a source=survey/yahoo/made_up tag
> still won't solve the issue as the source tag needs to be kept up to
> date, and you can link the source tag to the individual tag within a
> node (eg: the way is yahoo sourced, the name comes from the MMBW maps,
> a local marked it one way through local knowledge, and someone added
> speed humps using a GPS trace/waypoint setup.
>
> So the source tag contains a variety of sources, and no way of
> determining what applies to what.
>
> In the end, the crowdsourcing approach *should* sort out the wheat from
> the chaff eventually, but no doubt there will be some hiccups along
> the way.
>
> Matt
>
> Nick Hocking wrote:
> > "There is a substantial amount of data with no source tag. I'd really
> > like to
> > know whether it was traced or surveyed, because then I'd know whether
> > to go
> > down that street or not."
> >
> > Liz,
> >
> > I think there was a recent discussion about this and there was a
> > school of thought that said that if there
> > were public gps traces for a way then source=survey was unnecessary.
> >
> > I'm a serial offender for not adding source=survey tags but if the
> > general consensus is that it should be added
> > then I will manually add this tag to any ways that I have edited - I
> > could take a while though......
> >
> > I do upload public gps traces for all roads that I tag but I guess
> > that this is not a guarantee that the resultant way was
> > properly surveyed. E.G I have about 6 thousand miles of interstate
> > traces in USA that I have stopped
> > editing in because there is about to be another TIGER ravage.
> >
> > Eventually someone may use these traces without knowing if I was
> > actually driving on the road the whole time.
> > --------------------------------------------------------------------
Stay connected to the people that matter most with a smarter inbox. Take
a look http://au.docs.yahoo.com/mail/smarterinbox
_______________________________________________
Talk-au mailing list
Talk-au@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-au
