[Openstreetmap-dev] OSM-Editor compile report and first impressions

Nick Whitelegg Nick.Whitelegg at solent.ac.uk
Wed Sep 14 16:40:10 BST 2005




>Nick Whitelegg wrote:
>>>[osm-editor running quite slow on a 500mhz Pentium III with a 600 point
>>
>> track]
>>
>> PS if I get enough demand I can move the optimisation of osm-editor up
the
>> priority list.

>I think upload/download would be more useful - the slowest thing I find
>is waiting for the landsat image to update. An editor which can download
>from a GPS, allow you to mark up and annotate the track, then upload it
>to the server will be a real step forward.

The problem at the moment is that it grabs 4x the current window dimensions
- this is so that it doesn't grab new data every time you change your
position.

However, speeding this up could quite easily be done. If I added a new
keystroke to allow the user to move one window width horizontally, or one
window height vertically, and grab new Landsat data every time this was
done, then, as long as the window was not too big (say 400x400 pixels), it
should be as fast as OpenStreetMap itself.

>The easier it is for people to provide data, the more we'll get. (From
>that point of view, does anyone know if the editor will build on
>windows? Not really much of a concern to me, but if the editor was
>available to download then it may be the thing that tips people over the
>edge into helping.)

On the proverbial TODO list :-)
I agree, if osm-editor was a downloadable, click and run .EXE, it would
make it very easy for the end user. It would be possible to produce a
Windows version without too much hassle; the following steps would be
needed:

a) Convert from Qt3 to Qt4, as Qt4 is the first Free version available for
Windows. Probably not too much effort.

b) For GPS reading functionality, the jeeps library used to read the GPS
(which runs on Unix only) would need to be replaced by the modified jeeps
used in GPSBabel (which will work on Windows). Again, probably not too
difficult.

Once again, enough requests (even two or three would probably do...) and I
can prioritise a Windows conversion over some of the other osm-related
things I've got lined up. I do want to get the key/value stuff in OSM done
though, as hooking osm-editor up to OSM itself is rather dependent on that.

Nick







More information about the dev mailing list