No subject

[Darren Kriticos]

Hi Folks,
I've been forwarded a couple of postings regarding the use and choice of GIS packages for the analysis of specimen locality data for elucidating such factors as endemism and 'similarity'(?), and also for mapping the locality from where specimens were collected.

Gary Noonan made a good point about distinguishing between analytical functions and mapping functions.  HOWEVER, within any project or system there is rarely the need for only one of these classes of functions.  Usually, there is a need for different levels of sophistication in analysis and cartographic output.  In deciding upon which Desktop Mapping program (eg Mapinfo), or GIS to choose one should either consult a GIS professional or hit the GIS sourcebook.  Look closely at your own particular project/system and identify _inter alia_,
what functions you need to ba able to conduct;
what output requirements you have;
what inputs you need (e.g. base data);
what format those inputs are available in;
available vendor support;
available local 'guru' support within your organisation;
your budget.

Your final choice of package might be ARC/INFO, or it might be some transparent film, base topo maps and a set of letraset sheets.  Don't laugh, if your requirements are simple don't involve yourself with what will inevitably consume vast amounts of time and other resources.  On the other hand don't buy a package on factors like cost alone.  Many systems have been severely crippled by eager people buying a cheap package which wasn't able to cater to future needs and wasn't able to be expanded to cater for these needs due to a simplistic data model.  The managers of these systems were then left with the unenviable task of admitting to wasting a lot of money and re-investing in software, hardware, data conversion and training; and/or struggling with a system which was unable to cater to the new demands.

Generally speaking, in the academic areas of study of biological data, if you can't afford one of the 'big' packages then you'd be hard pressed to go past IDRISI.  It's exceedingly cheap and has one of the most sophisticated set of statistical routines for supporting ecology that I've come across.  It's drawbacks include the fact that its development has been very modular - like unix it's been put together by a succession of people.  The individual parts work well, but there is a problem with bringing it all together in an easy to use package - unix was founded on the motto 'user antagonism is good for the soul'.  The cartographic output is also a little basic.

Gotta go folks, hope the comments help.  Feel free to give me heaps.

ciao bene

Darren