[OSM-dev] Osmarender not always showing latest data

80n 80n80n at gmail.com
Sun Dec 14 19:09:00 GMT 2008


Now that there is starting to be a proliferation of sources of OSM data,
perhaps we should give some consideration to methods of verifying the
integrity of any source.

I'm wondering if perhaps we can compute some kind of rolling checksum that
can be compared between data sources.  It would then be quick and easy to
see if any server was ok just by comparing it's checksum to that of others
at any given minute.

BTW I'm not suggesting that the main api needs to generate this checksum
(although it would be great if it did), but rather just some way of
comparing one source against another - if there's a consensus then
everything is ok.

Anyone know a good, easy to implement algorithm, that could be used for
this?

80n

On Sun, Dec 14, 2008 at 6:00 PM, Rowland Shaw <rowland.shaw at gmail.com>wrote:

> All in that sort of time frame, so it looks like we've found our cause :)
>
>
> 2008/12/14  <milenko at king-nerd.com>:
> > This could be the fault of my ROMA server.  The landuse way is not
> present on
> > my server, although it does appear to be on the others and on the main
> API.
> >
> > I'm not sure how it's missing - my server must have somehow fallen out of
> sync
> > at some point.  Potlach lists that way as being created Nov 28th 2008,
> were the
> > other things you noticed missing also created around that time?
> >
> > -Jeremy
> >
> > Original Message -----------------------
> > One example would be the landuse=residential around Trimley St. Martin
> at:
> > It's (currently) on Mapnik at
> >
> http://www.openstreetmap.org/?lat=51.98797&lon=1.31644&zoom=15&layers=B000FTF
> > But not on Osmarender layer per
> >
> http://www.openstreetmap.org/?lat=51.98797&lon=1.31644&zoom=15&layers=0B00FTF
> >
> > And yes, the area is still in the data -- if you request a re-render,
> > I expect it'll reappear (I haven't).
> >
> >
> > 2008/12/14 80n <80n80n at gmail.com>:
> >> Rowland
> >> Can you make a couple of screen shots and describe what you see as being
> >> wrong.  Then someone might be able to investigate further.
> >>
> >> 80n
> >>
> >> On Sun, Dec 14, 2008 at 9:52 AM, Rowland Shaw <rowland.shaw at gmail.com>
> >> wrote:
> >>>
> >>> It's done it again since my previous message, too...
> >>>
> >>>
> >>> 2008/12/13 Rowland Shaw <rowland.shaw at gmail.com>:
> >>> > Every so often, the T at H tiles show some really odd combinations of
> >>> > data which show older versions of older nodes (for example, not
> >>> > showing access=permissive which was added to a way) whilst still
> >>> > showing newer ways that have been since added -- requesting a
> >>> > re-render through Informationfreeway has fixed it on the occasions
> >>> > I've noticed this.
> >>> >
> >>> > Primarily, I've noticed this in this sort of area:
> >>> >
> >>> >
> http://www.informationfreeway.org/?lat=52.02923827991905&lon=1.2023025518558124&zoom=16&layers=0B000F000F
> >>> >
> >>> > At first, I thought someone had reverted several hours of work, but
> it
> >>> > seems to be something else, that has now repeated itself two or three
> >>> > times (that I'm aware of)
> >>> >
> >>> > Any ideas of the cause?
> >>> >
> >>>
> >>> _______________________________________________
> >>> dev mailing list
> >>> dev at openstreetmap.org
> >>> http://lists.openstreetmap.org/listinfo/dev
> >>
> >>
> >
> > _______________________________________________
> > dev mailing list
> > dev at openstreetmap.org
> > http://lists.openstreetmap.org/listinfo/dev
> >
> >
> > _______________________________________________
> > dev mailing list
> > dev at openstreetmap.org
> > http://lists.openstreetmap.org/listinfo/dev
> >
>
> _______________________________________________
> dev mailing list
> dev at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20081214/6840e575/attachment.html>


More information about the dev mailing list