It seems to be an interesting tool, I'll give it a try.
For the 0.5 API issue you can adapt it to 0.6 using sed:
sed -i.bak "/version="0.5"/s/0.5/0.6/" srtm.osm
sed -i "/node/s/node/node version='1' timestamp='2011-09-17T12:00:00:Z'/" srtm.osm
sed -i "/<way /s/way/way version='1'timestamp='2011-09-17T12:00:00Z'/" srtm.osm
After the above command I could run osmosis 0.39 on srtm.osm

---Mensaje original---

Hi,

On Wed, Sep 14, Carlos Dávila wrote:

> I get the error below splitting SRTM of Spain My commands:
> mono Srtm2Osm.exe -bounds1 35.99 -9.501 43.79 4.6 -step 10 -cat 100 20 
> -large -corrxy 0.0005 0.0005 -o srtm_peninsula.osm
> java -Xmx2000m -jar ../../osmosis/osmosis-0.24.1-java5/osmosis.jar --rx 
> enableDateParsing=no srtm_peninsula.osm --bounding-polygon 
> file="/home/carlos/SIG/10m-admin-0-countries/spain.poly" --wx srtm_spain.osm
> java -Xmx2500M -jar /home/carlos/Paquetes/mkgmap/dist/splitter.jar 
> --mapid=65140001 --mixed=true --max-nodes=6500000 --description="SRTM 
> Spain" srtm_spain.osm

Meanwhile I'm useing phyghtmap instead of Srtm2Osm. Since I don't
merge the SRTM data with the normal OSM data, you can start with a
very low node id. As result, you can use all the current tools on
the data. And phyghtmap needs only a fraction of memory Srtm2Osm
needs for the same area.
Ok, phyghtmap, too, creates only v0.5 data. Haven't done anything
with python since several years, but to change to should be that
tricky. Only version and timestamp attributes needs to be added,
else osmosis is complaining.