|
|
| GeoCommunity Mailing List |
| |
| Mailing List Archives |
| Subject: | RE: GISList: Linux RS and GIS software for the masses |
| Date: |
06/10/2002 07:58:36 AM |
| From: |
Dimitri Rotow |
|
|
> GIS and RS. The idea of providing free software for these organizations is
> very appealing if they can use it to meet their needs without
> having to be a
> programmer or a Linux wiz. Within the next few months I expect to
> be able to
> invest the time to form my own opinion of the state of Linux-based open
> source GIS and RS software but I would value any opinions readers on this
> list might have regarding the readiness of Open Source GIS and remote
> sensing software for the masses. If it's not ready for the masses
> what would
> it take to get it to that point?
Let me rush in where angels fear to tread, at the risk of getting embroiled
in yet another Linux vs the world thread. Perhaps we can try some "thought
experiment" questions based on practical software development issues as seen
in real GIS software projects.
I take it your interest is in a GIS package that rivals the power and
progress of modern commercial GIS software. That's what I mean by a "GIS
application" in this thread: something at least as cool as a
state-of-the-art commercial offering. I assume Listers are not interested
in third rate applications that can sort of clunk along in a late 1980's or
early 1990's manner (setting aside for the moment the legacy interest in
ArcInfo..).
I think that conceptually you are getting ahead of yourself and asking a
very specific question when asking a broader question might be more
illuminating. Before zeroing in and thinking about open source / Linux GIS
applications, ask yourself the broader question about why there are no open
source / Linux applications to rival modern commercial software in other
applications areas, such as possible replacements for Microsoft Word,
AutoCAD or Adobe PhotoShop. The answer to that is closely related to the
answer to the question "Why is it that Linux is pervasive in server
applications but essentially nonexistant in interactive desktop
applications?"
I think there are three reasons: First, is not easy to assemble critical
mass for open source development so you have to have really huge numbers of
possible users to get a project going because only a small percentage have
the technical skills to do worthwhile development. If you don't have a large
population of interested parties, you can't recruit enough programmers to
work on the open source project to keep it moving forward. Second, it is
far more difficult to create applications with sophisticated user interfaces
than it is to create server appliances. Server projects can be accomplished
with a handful of people each working on separate modules. Interactive
applications require a much greater degree of teamwork and coordination,
with immensely greater demands for interdisciplinary work. Third, as
projects get more sophisticated (like interactive projects), it becomes more
and more difficult to make technical progress faster than the product
becomes obsolete. That tends to put an upper limit on the practical size of
open source projects.
All of these factors come together if you want to create an open source
interactive application for something like, say, word processing or GIS.
Given the smaller user population for GIS, it's not likely you'd be able to
assemble a critical mass for open development of a truly competitive
application. It is also not likely you could assemble the interdisciplinary
manpower in a concentrated enough form so the results of your work would not
be obsolete before you could finish your current edition. You could, of
course, create a consortium like the OpenGIS consortium and spend a few
years to hammer out something that was obsolete before you started, but I
assume that most Listers are interested in advancing the state of the art
and are not interested in spending a lot of time and effort to
institutionalize obsolescence.
To take a real life example, Manifold 5.00 Professional Edition is about a
million lines of code. Our spring and summer campaign to roll out
Enterprise Edition and the summer update to Professional has extended the
code base to about 1.4 million lines of code. Writing 400,000 lines of
complex code in less than a year is a very concentrated effort. Such
projects can only be brought to life in a coordinated way by large,
well-funded, dedicated teams of experts who work full time without
interruption, using very powerful machines for compilation and sophisticated
source code control tools. It's the only way to get things done so your
product is fresh and powerful upon release.
It's also the way people do things in Linux when the chips are down to
deliver fresh and powerful software. They put together commercial teams and
attack with the full sophistication of a well-funded, commercial effort.
They may base their efforts on "open" code, but ultimately what they sell
(yes... sell) is a commercial product. Tha
|
|
Sponsored by:
For information
regarding
advertising rates
Click Here!
|
