[GRASS-dev] GRASS-TNG

[Sören Gebbert]

Hi,
i agree with radek.

The core design of grass is very old. And so the code basis.
Implementing basic features
like thread safety -> parallel processing, a raster library which 
support 2d and 3d raster maps with a sophisticated time series support
will result in rewriting the core of grass.

So i think it is a good approach to start from scratch using the 20 
years of great experience with grass to create a successor. The software 
technics have radically changed in the last 20 years and so the 
hardware. And IMHO to profit from those changes, we need to start from 
scratch.

This also allows us to think about new great features, which are even 
not available in modern commercial GIS.

So lets collect new ideas at
http://grass-tng.no-ip.org.
This project will only work with the support of the grass community.
Take this as chance to discuss a new and better GIS core design.
All points of view are important, the views of the users which kind 
features they expect, the views of the developers which features are 
possible to implement and which not.

But important to know, we currently discuss the core design of 
GRASS-TNG, not the features of hundrets of grass modules. :)

Best regards
Soeren

Radek Bartoň schrieb:
> On Wednesday 21 of February 2007 15:19:55 Paolo Cavallini wrote:
>> So I assume it is a fork, which is generally a Bad Thing.
>> May I ask you why forking, and which development forces do you have to
>> carry on what appears a daunting task?
>> All the best.
>> pc
>>
> 
> If it is fork depends on what is definition of fork. Generally it should be 
> independent project which would provide modern programming framework for 
> analytical modules similar to that in current GRASS. So any user which is 
> used to work with GRASS command line would be familiar with new system but 
> inside it'will use completely new design oriented to OOP, extensibility, 
> parallelity and dynamic languages. Theoretically it should be compatible with 
> any GRASS's GUI developed over modules in the future.
> 
> IMHO a current state of core parts of GRASS is so unogranizes and oldstyled 
> that any progressive development is very difficult. That is why I think that 
> start from scratch and only  take good ideas from GRASS is now the best 
> solution how to make a 21th century open-source GIS realizable.
> 
> For next one year I'll be working on making its design and prototype 
> implementation as my diploma work so even if this project wouldn't keep up it 
> would be at least a research of GIS domain. I have spoken with a few current 
> GRASS developers and they invite my ideas so I wish it won't happen.
> 
> Only support from comunity I miss for now are discussion of ideas and needs of 
> features but any kind of help or invention is welcomed.
>