[Thuban-list] future of Thuban

[Jan-Oliver Wagner]

Hi,

On Wed, Apr 02, 2003 at 01:51:51PM +0800, Yao Heling wrote:
> 1. shouldn't there be a select tool? All I can come up with now is to
> used the identify tool, which brings up the dialog, which users may not
> need;

there are already internal improvements to support more sophisticated
ways for selection. Most notably that it is possible to select
more than one object. Eventually this will be available through
the user interface as well, but I can not promise when we have
the time for it (there are some other tasks currently with a
higher priority).

However, it would be interesting in how you would like to function
Thuban regarding selection?

> 2. I remeber there were talkings about the use of ogr instead of
> shapelib on the list, how' that going? As I plan to learn/use OGR
> myself, I may be of some help if ogr is used;

It is on the mid-term wishlist, but currently we have no free resources
(or financial support) to work on this.
However, Martin works on extending Thuban to access spatial databases
according to the OGC specification.
This will also refine the concept of different layer types which
should make it more easy to add OGR support.

> 3. As I understand, thuban is GIS viewer as describled on the website;
> but I believe an advanced viewer would have some basic spatial analysis
> capabilities incoporated, is that planned for thuban? at least the
> infrastructure should be there for users to add needed functionalities,
> such as buffer analysis, overlay analysis etc.

I am not sure whether we should make Thuban a full GIS. It is actually
designed to be a viewer/explorer. Interfacing with other tools
such as GRASS might be more interesting since there is a lot
development power.

What we have in mind is that it is possible to build specific GIS
tools based on Thuban. We have done this already for one
tool, GREAT-ER (greater.intevation.org).

Also, it is easy to extend Thuban with some extra-functionality
(you will find examples in the CVS). Buffering etc. should
be possible with this mechanism.

> 4. When I've selected a polygon in thuban, everything (zoom, pan etc.)
> becomes extremely slow; is it because Python is not up to the task; an
> algorithm issue? maybe we should consider using numpy?

I've not experienced a serious slow-down so far, but maybe you layer is
much bigger than what I use. So far, we have not paid much attention
to performance optimzation. There are definitely some screws to
improve the situation.

> 5. If thuban is to use ogr and numpy, it would look like openev a lot?
> maybe some ideas could be borrowed from openev?

OpenEV is basically a library with a sample viewer application.
But you are right, both tools seem to share quite a lot.

> 6. license issue: since thuban's license is GPL, some users may feel it
> too restricted; is an license change to LGPL possible? (openev is LGPL)

OpenEV seems to be developed by Atlantis to derive proprietary products.
We do not intend to do the same. We'd rather like to see
Thuban-based GPL tools in the future.
Basically, GPL is only restricting the right to restrict rights.
Commercial success exclusively based on Free Software is possible.
I hope to convince more people :-)

Best

	Jan
-- 
Jan-Oliver Wagner               http://intevation.de/~jan/

Intevation GmbH	              	     http://intevation.de/
FreeGIS	                               http://freegis.org/