[josm-dev] Audio mapping Plugin

[Brian Schimmel]

Hi everybody,

I'm writing a plugin for JOSM to enable more efficient audio mapping. It's
still at a very early stage, has a bad usability, but works good for me. I'm
now planing to put this to the next step, publishing it and making it
accessible to other testers and coders. I've got some questions concerning
this. Before I ask them, let me explain what I think audio mapping should be
like:

The user may use any audio recording device he likes. Depending on his
situation he may do one large recording for his mapping tour (as I do, up to
three hours in one file) or record some small snippets when he needs to. It
is assumed that there will be some way to synchronize the audio time to the
GPS time. By that it is possible to show the location where the sound was
recorded. Of course this is not a single location but a subset of the gpx
trail, with a cursor moving along the GPX trail while the audio plays.

Synchronized audio mapping allows some new methods to map: For example you
drive along a road, and each time there is a side road connecting, you say
"A road from the left, called Huberstaße, connects HERE from 8 o'clock.",
saying "HERE" at the very point where the roads meet. Don't get me wrong, 8
o'clock is a direction, not a time :) I've done this hundreds of times now,
and it really gives sufficient accuracy if you do the synchronization
properly.

While digital photos will have EXIF data with time stamps, audio recordings
mostly have no time meta data at all, or have them saved in many different
ways. I've found some ways to syncronize any audio, but this is beyond the
scope of this mail.

Currently, I did not implement a real plugin. I used
org.openstreetmap.josm.gui.layer.GeoImageLayer, copied it to a file named
GeoSoundLayer.java and changed what was needed to change.

Now you know what I'm doing, here's my question:
How should I go on? Shall I restructure my code to make a proper plugin, so
that there are no changes to the JOSM core? At first this seems to be the
only good way.

But on the other hand, there are many similarities between photo mapping and
audio mapping. So I might take all their common features, put then into a
new class, and subclass the photo layer and the audio layer from that. I
know of some people who take text notes which could also be geotagged by
their timestamps. This seems to be the even better way to get nice code.

I've recently seen that there is a package markerlayer which resambles this,
and it already has ImageMarker and AudioMarker. But GeoImageLayer has
nothing to do with this, has it? I did not have time yet to look deeper into
this package. Is there any policy to put together the geoXxxLayers with
MarkerLayer, or are this two very different things?

At the moment, my code isn't nice, but it works and should not break
anything else. Would it be wise to commit that to the SVN now?

mfg Brian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/josm-dev/attachments/20080201/456b0ce3/attachment.html>