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

nick at hogweed.org nick at hogweed.org
Thu May 25 20:24:17 BST 2006


Oops... didn't mean to send that last email! Damn those keyboard shortcuts :-)

Anyway....

----

On Thursday 25 May 2006 15:27, Ben Gimpert wrote:

> I've just cut what was going to be a long-winded, overly-defensive,
> flame-bait critique of Java.  Instead I'll attempt my appeal with some
> bullet points:

Didn't think this thread would stir up controversy :-) but to summarise
what I said in another part of the thread, I've been a little unhappy lately 
with the productivity C++ has given me, particularly when working in parallel 
with C++ (osmeditor2) and PHP (the Easter version of the renderer)
so with gcj going native, and the fact I have good knowledge of Java, I 
thought it might be an idea to switch. To go into a language-rant of my 
own.... Though Qt is IMO an excellent, well-designed and elegant library, 
I've never really got on with the design and philosophy of "Standard C++" - 
I've always found it a bit hacked-together, and the fact that objects are not 
pointers/references by default is a recipe for bugs.

The idea wasn't so much a 'greenfield rewrite' it was basically keep the same 
design, but use Java instead.  I anticipated it would take no more than two 
12-hour coding days, and the resulting increased productivity and 
code-sharing with other editors would benefit all. However I am by no means a 
Java-worshipper and am happy to use any language which will give me 
productivity with osmeditor2 and allow me to do what I want with it. As I 
said in another threat, I would like Linux users to be able to compile (if 
applicable) and run without dependencies, Windows users to be able to 
point-and-click on an EXE and go, no need to use gpsbabel or any other piece 
of software. From those points of view, Ruby may well be a wise choice :-) 
However it would require research.

To get a bit more philosophical I'd like to respond further to one or two 
points which I didn't previously. This might be purely irrational, but I 
would not want osmeditor2 to be abandoned in favour of the applet or JOSM.  
I'm sure everyone on this list would agree, but writing a piece of software 
is a creative thing, and to abandon it after 12 months of work, even if 
pragmatically this might have its advantages, is something I'm not terribly 
happy on doing. I only wish I could spend more time on osmeditor2, to make it 
more "professional" but sadly time is very much at a premium at the moment. 

Anyway, enough of that, back to coding :-)

Nick






More information about the dev mailing list