Thanks!

I did not see any profile when I tried just a few minutes ago. Did I miss the show?

Sjors

First reactions are very enthusiastic. Good work!

So far, I am focusing on walking and cycle routes. I have tried a few random examples and those were less than 100 nodes. That said, we should be able to figure out a way to deal with tens of thousands of nodes.

Perhaps, as a start, you could fetch the altitude profile only for routes with less than 500 nodes? I am reasonably confident my home computer can handle that a couple of times per minute.

And then we need to start thinking about those bigger routes. I have a couple of thoughts here:

* the altitude profile is only a small figure; 100 evenly spaced altitude samples is enough for any route, for most landscapes

* my server could process a big input and pick 100 evenly spaced samples; that should seriously reduce its work load

* sending a 30.000 waypoint route over the network only to use 100 points is a not very efficient use of bandwidth. But requiring route services to do their own filtering adds complexity.

Let’s take it one step at a time though. Do you need any help? I am not sure which language you are using, but if there is a protocol buffer implementation for it, that would be the most efficient solution.

The longest route that has been generated so far comprises over 30.000 nodes and many routes are over 8.000 nodes, is your server/application up to it to handle such render request multiple times per minute?

[1] http://tile.openstreetmap.nl/~lambertus/routing-world/index.html

Here is the ‘Subscribe to Comments’ plugin for WordPress: http://wordpress.org/extend/plugins/subscribe-to-comments/

After lots of reading, I have eventually decided (decision making gets easier when you are hungry) to attempt a REST-full approach.

The first reason is aesthetics; I like REST because people say it is more elegant for various reasons.

The second reason is that it is stateless, which makes it easier to host on something like the Google App Engine, which is also stateless. My intuitive guess is that if Google thinks stateless is the way to go for their massive systems, they probably had some smart people think about it. And OpenStreetMap is pretty big already and will probably become massive.

I just found out that the Google App Engine is available for everyone now [1]. Would have been nice if they emailed people on the waiting list that piece of information.

I am going to try to port my application to the App Engine, just using a fraction of the data to see how quickly I cross the 500 MB storage limit.

They have also released some information about pricing. I should be able to calculate how much storing the entire SRTM data set would cost, etc.

Fun times awaiting!

[1] http://googleappengine.blogspot.com/

With XML-RPC/SOAP the focus is on methods (verbs), with REST the focus is on objects (nouns). There are a couple of advantages to REST, but I’m not sure if they apply in your case. It might be nice reading about the differences. I always liked How I Explained REST to My Wife,