[Roadster] state of the roadster
arenner_dev at wowway.com
Mon Jul 17 19:00:00 PDT 2006
Ian McIntosh wrote:
> Hi Everyone,
> Update time.
> First, personal: I am still in South America, but perhaps more settled
> now. I would very much like to continue working on Roadster.
> Up until this point, Roadster has had only a few people working on it.
> If Roadster is going to be a success, it'll take many more people
> So, this is a call for help! This mailing list has 70+ subscribers now,
> and I'm hoping there are some people here who want to get involved.
> In addition to bringing more people in, I'd like to reduce the scope of
> the project for version 1.0.
> Version 1.0 feature proposal
> 1) move map data loading (and rendering?) functionality to an external
> library (libroadster?)
> 2) external library will use a better (faster/smaller) data format (no
> more MySQL GIS) stored in a SQLite database
> 3) ability to zoom from street level to state/country level *quickly*
> (which probably means storing data at several levels-of-detail)
> 4) a new GUI in a high level language (C#, Ruby, Python?) with only
> basic map navigation functionality (mouse drag, edge scroll, zoom
> buttons, zoom rect tool, history back/forward buttons)
> I see this as a solid base for more advanced functionality, and we cand
> decide later where it should be added (in the application, in the
> library, or as an application plugin).
> At that point, multiple frontends could be developed independently using
> native toolkits (GTK+, QT, Cocoa, Win32).
> So that's the vision. What do you think??
> roadster mailing list
> roadster at cairographics.org
I have been messing off-and-on with databasing the Tiger/Line data in
different ways for almost two years. I like MySQL. I have also tried
Gigabase and sqlite. Gigabase was just too difficult to use and I feel
that it would hinder a projects ability to attract new developers as
opposed to a SQL database that everyone is familiar with. Sqlite has too
few features. Spatial indexing (of some form) is a must-have. Full-text
searching is also really nice to have.
I believe that you could completely seperate the database from the
mapping applications. This would help adoption because many people might
be interested in web-applictions, others might be interested in
hand-held computer software.
I highly recommend anything in Python whenever possible. That is what I
usually work in.
I am willing to help, but my interests, given my limited time that I can
work, would be primarily in supporting applications. For instance, I am
toying with the idea of a python-glade application that allows the user
to download data from yahoo or google local for a limited area and at
the data to the database.
More information about the roadster