|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | RE: GISList: Re: neverending story was: Automatic data recording |
| Date: |
07/03/2003 02:45:00 PM |
| From: |
Dimitri Rotow |
|
|
>
> though, let's go back to the GIS...
>
> >And then again we return to the question of GIS use. There really isn't
> >anything in GIS for Linux that is remotely comparable to what a mere $250
> >buys you in commercial products. The main reason is that a modern, high
> >quality, fully functional professional GIS that can handle vector, raster
> >and terrain elevation data simultaneously with elaborate support for
> >sophisticated GIS tasks and also includes Internet map serving
> capability is a huge project.
> >
> ---------------------------------------------------
> I am sure you know GRASS http://grass.itc.it/ and plenty of add-ons for
> it for virtualy any purpose you may need for your GIS projects. Also a
Please, be serious. GRASS is antiquated raster-oriented GIS and does not
remotely contain the feature set most professional users expect in a modern
GIS. I doubt you would ever convince, say, most companies running ArcGIS,
Manifold, Intergraph or even MapInfo to use GRASS. Well, perhaps you could
find a few whose socialist ideology transcends their common sense to do so,
but I don't think you could find many well-run organizations that would
agree GRASS is as functional a modern GIS as, say, ArcGIS or Manifold.
> freeGIS http://freegis.org/index.en.html project. There are also many
> free single-purpose or specialized GIS packages (also for MS Windows) like
> Microdem http://www.usna.edu/Users/oceano/pguth/website/microdem.htm ,
> 3DEM http://www.visualizationsoftware.com/3dem.html,
> fractal landscape simmulator-Terragen
http://www.planetside.co.uk/terragen/,
SPRING GIS http://www.dpi.inpe.br/spring/english/
that emerged as results of research and development activities in the
universities.
Yes, and for the most part (but not always) they look like disparate R&D
tools and not like polished commercial packages with up-to-date Windows
interfaces: for that matter, what most people want are a highly integrated,
highly functional single package so they don't have to hassle with learning
20 different bits and pieces to accomplish what they want to do. For
example, Microdem is an excellent package but it only does a small part of
what people want to do with professional GIS. I think if you are seriously
suggesting that the above set of bits and pieces could in any realistic way
be used to replace, say, Intergraph or ArcGIS 8, you are not very familiar
with what people do with such packages or why they buy them.
>It requires the concentration of sufficient technical and
>financial resources to create about a million lines of code per year (a
rate that is necessary to be sure the product does not become obsolete
during the time it takes to code it). Note that the task is not to
painfully assemble a total of one and a half million lines of code over the
course of ten years, because then the result is obsolete when complete. The
task is to sustain huge, focussed development throughput that can virtually
move mountains within a single year. Otherwise, one cannot advance the
state of the art in GIS.
>
----------------------------------------------------
I do not think 1 000 000 lines/year is a measure of quality. It is the
algorithms that make the software run. One can write 100 rows for the
task that is solvable with 20. Also there are algorithms that take 1
minute to do what is possible to accomplish in 10 sec. Moreover -
a huge amount of these rows are produced automaticaly by any favorite
"VisualSomething" to produce a GUI. While a "real algorithmic GIS" part
of it occupies just a relatively small part of the binary space.
Am I using "obsolete" GIS software ?
Well, is it important if it can easily accomplish everything I need ?...
-----------------------------------------------------
You are missing the point. Most professional GIS users have a very large set
of requirements and desiderata, and many amateur users like the quality,
convenience and power of using professional tools as well. Algorithms are
not usually the labor-intensive part of modern, sophisticated products. It
is the seeming infinity of small details that users want the package to take
care of that require a lot of coding.
I could use, say, ArcGIS as an example but I am more familiar with Manifold,
so I'll use that as an example. In any case, the point that a
sophisticated, commercial GIS provides a highly feature-rich environment in
response to user needs that requires the coding of millions of lines of
code - that is the point regardless of what GIS is used as an example:
For example, Manifold imports about 80 different formats, many of which
involve various options. All that stuff has to be coded to do a good job
for the user, requiring a vast amount of code. To take another example, the
modern way is to integrate an Internet Map Server (IMS
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
