[OSM-talk] Server slowness

[Nick Black]

On 1/15/07, Kristian Thy <thy at 42.dk> wrote:
> On Mon, Jan 15, Nick Black wrote:
> >> PostGIS, at least (I don't know much about MySQL's spatial abilities) has
> >> GiST indexing, which is a nested polygon bounding-box index. If you have
> >> geodata, why do you not store it in a geodatabase and get all the bells
> >> and whistles for free, instead of trying to come up with a home-grown
> >> solution?
> >
> > Because it is more efficient, simpler and fulfills the tasks that OSM
> > is trying to accomplish.
>
> Yes, that is obvious from the performance of the server over the last
> week. It is also obvious from the threads you link that you are totally
> in control of the situation, and I will refrain from trying to help you.


Please dont take my comments the wrong way - you ask why OSM dosnt use
a geodatabase and I give you an answer.  No-one is claiming that the
server is lightening fast, or even not slow, and I am certainly not
claiming to be in control of any situation - least of all the OSM
database.    We need to need to know exactly what is causing the slow
responses over the last few days and then see how using a spatial
databse would effect these situations.  I use PostGIS every day - its
great for storing, managing importing and exporting spatial data.  But
its not a panacea.

>
> > Sorry, I dont really follow this point.  OSM XML would still be needed
> > to distribute data from the API, Planet, etc.  I dont think that the
> > XML structure (itself always subject to reconsideration) forces the
> > data structure.  Using  OSGEO Linestrings,  Multilinestrings however,
> > would force the data-structure.
>
> It's geometry, for $DEITY's sake. With attributes. There's nothing new
> or special about the OSM data - it's the surveying that's the strong
> point of the project. I see nothing in the OSM data that cannot be
> portrayed with polylines, points and polygons.
>
> > "Anyone who thinks the necessary technology is already out there to make
> > OSM more efficient ought to prove it."[1]
> >
> > [1] -
> > http://lists.openstreetmap.org/pipermail/dev/2006-September/002119.html
>
> "I'm strapped to the backseat, so you can blame me for being a backseat
>  driver all you want.  It's quite frustrating."[1]

If you feel this way, then what do you need to be unstrapped?  OSM
should allow equal contributions from everyone - if there are
impediments to your full participation in the project, let us know.


Nick
>
> [1] -
> http://lists.openstreetmap.org/pipermail/dev/2006-September/002120.html
>
> Anyway, since this is a discussion that seems to crop up from time to
> time, perhaps Nick Hill's post you quoted should feature quite
> prominently on the "Get involved" page on the wiki.

Agreed - I'm assembling the results of previous tests and discussions
together to make a wiki page.

Nick