[OSM-dev] Re: osmeditor2 to Java, and a common Java OSM client library (was:Re: [OSM-talk] Java and freedom)

[Nick Whitelegg]

Sent by:        talk-bounces at openstreetmap.org
To:     talk at openstreetmap.org
cc:      
Subject:        Re: osmeditor2 to Java, and a common Java OSM client 
library (was:Re: [OSM-talk] Java and    freedom)


>Offline editors are pointless given the current state of the OSM API.
>Until the *API* (and not each and every client) deals effectively with
>versioning and merging data, then offline clients are Broken As
>Designed.  The nicer the offline client, the longer you will spend
>using it, and the more likely conflicts between data will occur.

osmeditor2 is not strictly speaking offline though. Offline editing 
followed by batch upload is not possible in the current version of 
osmeditor2, as the mechanism I used for batch upload was no longer 
possible when we moved to the 0.3 API. osmeditor2 is predominantly an 
online, 'live' editor at the moment.


>Nick, I understand your desire to put your efforts into an all-in-one
>editor that also talks to your GPS, but if you go with the 'small
>pieces loosely joined' mentality that got the web so far, and let GPS
>babel handle the GPS device, why would you write Yet Another Java
>Editor and not just hack on the applet or JOSM?

>From a usability point of view, I prefer an all-in-one solution. Given 
that there is a lot of existing Java code out there, and that gcj means 
that a lot of my original issues with Java are less relevant, I thought it 
might be a sound idea - particularly given that (IMX) debugging is easier 
in Java than in C++ so it would have long-term maintainability advantages. 
I don't really care what language I use though, to be honest, as long as I 
can develop an editor which aims to meet my needs and the particular 
target user base I'm aiming at (outdoor users such as walkers, 
particularly in the UK). This has not yet been achieved with osmeditor2, 
but one thing I really want out of it, whatever technology I choose, 
whether that be stay with C++ or convert to some other language, is 
point-and-click simplicity on Windows along with easily compilation on 
Linux, no reliance on third-party tools.

>If anyone wants encouragement to do anything in particular, I reckon
>that the priorities are working on the website, online editors, API
>and in particular getting edit history and talk pages for the maps.
>Oh, and I think that putting SVG output online so that people can see
>the fruit of their efforts will make the biggest contribution to OSM
>in the near future.

How does working on the renderer fit in here? :-) Think that's possibly 
amongst the most important of all the required tasks, hence my 
prioritisation on it ATM.

Nick