[OSM-talk] The long tail

[Tom Carden]

On 06/07/06, Immanuel Scholz <immanuel.scholz at gmx.de> wrote:
>
> I need different parts of the database at different time. As a coder, it
> is mostly needed for doing offline testing and performance measurement.
>

I had hoped for this too.  Maybe Steve could get some NDAs for people
wanting to experiment with the live database (a slave of it, I mean) -
to spread the liability pre-OSMF?

> To write on history stuff, a history in the dump would be nice. To check
> for concurent modification issues, the timestamp is necessary. To work on
> GPS traces, these table is helpfull... you get the point.

Some of these things are coming.  I don't see a problem with
timestamps on the data - at least marking the time the export was
performed would be very sensible.

>
> I just don't want to beg every time again to get the next piece of data. I
> have to beg customers in bussiness each day, but never in open source
> projects dedicated to distribute data.
>

What others, apart from wikipedia, are relevant examples?  I'm very curious.

>
> The other reason I lately insists in getting the database is Steve's
> strange reasons suddenly popping up from nowhere the same time his old
> reasons were thinned out by faster scripts and people declaring their lack
> of privacy interest.
>

They don't pop up from nowhere - privacy has been the main reason for
not doing a database dump for ages.  When Steve finally did do a
limited export (the first planet.osm) it was the first time he'd
tried, and so reasons of performance then entered into the equation.
Nothing funny is going on.  There may be differences over opinion
about the legality and ethics of certain actions, but nobody is
playing with you.  These differences in opinion can be resolved in
time, preferably using the clean-room/paranoid approach that keeps OSM
safe, but once there is an OSMF then depending on member wishes and
who is in control... feel free to be reckless.

Best,

Tom.