I laid out my case for integrating GPS, Cameras, and Blogs into a tighter package better for travel blogging. The first project to come from that is one to convert GPS waypoints into blog items. I've actually written a very primitive script to do just that. Now it's time for some planning.
HIGH LEVEL RUBBISH
====================
The basic tool is to save waypoints from a GPS to a file and then process that file, breaking up each waypoint into its own article inside a desginated blog.
OziExplorer Waypoint Processing
I hate Garmin and especially their lousy, proprietary software. They really suck. Therefore, I am not even going to worry about processing their stupid .mps track and waypoint files. Fuck them. People should find alternatives to Garmin.
Intead, I will focus on processing OziExplorer format files. While OziExplorer is sort of weird, it works and is semi-open.
Turning Waypoint Data into a Useful Article
There's not that much useful telemetry in a waypoint. Generally you'll just find a short, unfriendly tag name along with latitude and longitude (in decimal degree format). Possibly you'll get the date the waypoint was set, an altitude, and a description, but don't count on these being useful fields.
So what to do with this telemetry you've turned into a blog?
1) Make it a geourl anchor so that other geosearches can return this waypoint
2) Make it downloadable as a gps waypoint file so that others can use this waypoint in their own gps
3) Convert it into a small map inside the blog so that readers can see the area we're referring to.
.pl versus XMLRPC
Presently the code is just a perl script sitting on the same server as MoveableType. I execute the script from an ssh command line and it directly calls the MoveableType APIs.
This is an easy way to do it at first. There's no need for gross web-based interaces, or for learning XMLRPC.
In the long-term, though, it's not the best solution. Why?
1) It only works with the MoveableType API library. Forget Blogger, LiveJournal, or any of the others
2) It requires direct, root access to the server to execute the scripts on the data (which also is local)
What might be better solutions?
1) Local client that talks to the blog server via an XMLRPC interface.
You'd have a little program, gp-it.exe, that sat on any machine you wanted, would take the waypoint data and process it locally into XMLRPC instructions for the web server. Thus you never have to have direct access to the blog's server. That makes it safer, tidier, and allows you to post GPS-blogs to any blogging software that supports the XMLRPC interface.
2) A GPS-Blog pane in the administrative interface of MoveableType
Via the admin interface, you'd upload the waypoint information, configure the output, and process it. The advantage is that you don't need some local client software. (makes it easier to use internet cafes, for instance) It also provides a safer, tidier interface to the server. But like the existing perl-script option, it restricts you to working only with the MoveableType API.
TO-DO LISTS
=========
1) Check on making the blog auto-ping the GeoUrl add page. Check with joshua-geourl@burri.to
2) Link to MapQuest as a short-term inline map alternative. [Done, but not inline... opens to external window]
3) Find something inline map alternatives, like IndyJunior
4) Auto-generate "download this waypoint" files
5) Prevent waypoint/blogitem duplication
6) More robust processing regexps. incl strippping out date and altitude.
7) good subroutines for converting lat/long formats and tdatetime formats.
8) Give it a destination category.