[Freegis-list] Re: GIS grant to help map cities worldwide

[Adrian Custer]

Hello Jan-Oliver,

I'm a bit puzzled by the cross-post you sent to the FreeGIS list since
it started in the midstream of the exchange. Here, is a quick take on
the some issues in the use of GIS for spatial research in agricultural
systems. It may be helpful to Mr. Roger Longhorn. These issues have no
relevance to the financial costs of the software but rather pertain to
the control of the tool. Could you please forward it to him if he is not
on the freegis list?


While there are many arguments in favour of *free* software, the ability
to modify the source code is critical for any research using geographic
information systems. If Mr. Longhorn and the CGIAR intend to use GIS for
spatial research in agricultural systems, success will only come from
the long term accumulation of useful tools, routines and data. Contrary
to what the commercial vendors repeat, our understanding of spatial
analysis is not nearly sufficient to be able to rely on pre-packaged
algorithms for analysis. The major costs to the consortium will be
involved in the time required of the research scientists to develop
useful models, build a coherent set of baseline data and develop useful
analytic algorithms. This investment will only be recouped by the
consortium if, at the end of the process, it has a tool that it can then
use in perpetuity. The difficulty in getting an estimate of the growth
of cassava across Africa, for example, does not lie in mapping Africa
but in getting a reliable set of weather estimates, a reasonable plant
growth model and a decent estimate of the areas where cassava is grown.
By all means, CGIAR should pay for software to help shorten their time
to completion, but the consortium should buy *free* software and pay for
consultants to help the consortium's scientists program the models,
acquire data and understand the issues involved in spatial analysis.
These consultants, and the product of their labour, is the key to the
successful use by CGIAR of GIS for research. Even the best spatial
ecologists don't understand the robustness problems involved in the
manipulation of several data layers, ignore the issues of error
propagation and misunderstand the applicability of the statistical tools
in spatial analysis. Purchase advice and help, not licenses. The
alternative will be a slew of projects, each having its analysis
assembled piecemeal and no simplicity gained over the next few decades
of work. 

On the other hand, if the CGIAR is using GIS to map current situations
for funding proposals or to illustrate political statements, current
proprietary software is probably sufficient, quick and easy. However,
once again the product of the labour of the researchers of the
consortium will only be partly available to the CGIAR. There will be no
way, for example, to share with government agencies or university
research groups around the world the analytic tools with which the CGIAR
reached its conclusions. 

The best solution for the CGIAR would be to join the FAO and several
other groups, allocate a source of funds over the next five years and
pay someone to assemble one, or several complementary, systems to serve
their needs. At the end of the process, the consortium would have a
system that it owned, that it could modify and that was targeted to its
particular needs. 

Good luck in whatever path you choose, 

adrian custer
dept. of entology
u.c. berkeley




On Sat, 2003-02-08 at 09:25, Jan-Oliver Wagner wrote:
> [ again crossposting EGIP and FreeGIS ]
> 
> On Sat, Feb 08, 2003 at 04:50:38PM +0100, Roger Longhorn wrote:
> > Even "free" software has a cost.
> 
> Absolutely correct. And it should never be seen in another way.
> 
> Free Software is about Freedom, not costs.
> Free Software is definitely assotiated with costs (as is any IT activity).
> But you have the freedom to choose the type and timing for the costs.
> 
> > While Jan has a point, free (open source) software is not always the 
> > answer. Its use depends upon the application and the skill level (both in 
> > IT and in the application sector) of the user, plus the level of complexity 
> > in spatial analysis that is actually needed by the user.
> 
> You are right that costs are associated with the tasks to perform.
> They usually are not with the license fees.
> But Free Software _is_ the answer here because you have the freedom
> to take advantage of your skill level or pay otherwise.
> With proprietary product you only have the latter option.
> 
> > While completing some IPR-related "spatial information and tools" (read GIS 
> > software) research for CGIAR - the Consultative Group for International 
> > Agricultural Research - in which we specifically looked at open source GIS 
> > and related spatial information management tools, we found that most of the 
> > CGIAR's 16 international institutes relied on commercial GIS packages 
> > because of the level of support available, at various levels.
> 
> Note that commercial is not the opposite of Free Software.
> Both, proprietary software and Free Software can either be commercial
> or gratis.
> Professional support for Free Software (and yes for GIS too) is
> available. You just have to pay for it. You get the same
> gurantees as for proprietary products (software is always just technology),
> but usually a far better service (because you are not bound to a
> specific vendor, the contracted service company does the best possible
> service for you to keep you as a client).
> 
> > The support they needed ranged from advising on hardware and operating 
> > system software, to actual implementation of the software, use of the 
> > software with the specific application(s), then on-going support and 
> > software upgrades. The problems faced in the spatial info world are 
> > legendary, as we have vastly differing sources of spatial data, different 
> > formats, different scales, different applications and goals/objectives in 
> > using the GIS tools. Complicating this are the requirement for many 
> > research institutes to share, integrate and use their spatial data on a 
> > global basis with sister organisations and/or other nationally-based 
> > research institutes.
> 
> All arguments to prefer Free Software over proprietary.
> 
> > While OGC's excellent work on interoperability is going a long way to 
> > assisting in this latter respect for commercially available GIS software, I 
> > am not informed as to the degree to which the open source GIS community is 
> > taking the new OGC Web-related specifications on-board.
> 
> Note first that OGC derived from the former GRASS committee :-)
> Second, yes Free Software adopts standards. Usually faster
> and better than others if the standards are sensible. Look
> at Deegree, MapServer, PostGIS for OGC standards if you like.
> 
> > Just to make clear that I am totally in favour of *appropriate* open source 
> > software, I acquired my new laptop a few months ago, which came with 
> > Windows XP installed plus Microsoft Works, but not Microsoft Office, of 
> > which I have a legal copy. However, my copy of Office 2000 (not 2002) 
> > requires an initial load of the basic package, then loading of three sets 
> > of (freely available) service pack upgrades - which takes about an hour! 
> > Either that, or purchase an upgrade to Office 2002 for a couple of hundred 
> > Euro - and then still have to install service pack upgrades.
> > 
> > Instead, I bit the bullet, as we say, and installed a completely free copy 
> > of Open Office 1.0 from the free CD-ROM on a computer magazine. It has been 
> > running so well - and with such good Office compatibility - that I have had 
> > no reason to load and update Office 2000 on my system, let alone upgrade to 
> > Office 2002. You can buy fully supported Open Office implementations 
> > (immediate on-line support) for as little as 70 - 80 Euro, with support 
> > from such firms as IBM and Sun Microsystems.
> > 
> > This was *not* the experience I had two years ago when trying to install 
> > and use GRASS on my desktop system. I never got fully to grips with GRASS 
> > and in the end reverted to copies of other "low-end" GIS software. The fact 
> > that there are several versions (upgrades) of GRASS available did not help, 
> > as one source of support might be quite good for one version but not for 
> > later (or earlier!) versions.
> > 
> > Fortunately, several of supposedly "low end" GIS suites (i.e. not ESRI, 
> > Intergraph, Autodesk or MapInfo products) are now very powerful and for 
> > many applications, especially "one off" applications, these fully supported 
> > GIS packages are more than adequate, for very little cost.
> 
> I merely hope you understand low-end as
> low-cost-to-get-the-software-onto-your-harddisk and not as low-quality.
> But again: Costs are assoiated with the task as such, not with use-licenses
> or the costs to just have a software running on you computer.
> Think of the total cost of operation.
> 
> > More importantly, to my mind, is the *data* issues, not the GIS software 
> > issue. The announcement states that the grants for software will help "to 
> > collect and analyze data about basic infrastructure and the current status 
> > of housing." The costs to do the data collection, whether in terms of 
> > finances or human resources, will *far* outstrip the cost of the GIS 
> > package that will be used at the end of the process. Once that data is 
> > collected, will it be made freely (or cheaply) available to other users in 
> > these same developing countries, to perhaps help analyze other problems? 
> > Who will pay for the data maintenance and update, the quality control, etc.?
> 
> Indeed the data are important. However, strong proprietary companies
> like Esri have the tendency to make the users data turn to proprietary
> formats difficult to migrate at a later stage. There is also
> the strong tendency to move the users into a proprietary working
> environment they can't leave except at very high costs (note, that
> it is the usual strategy of proprietary vendors to set the update
> costs just a little bit lower than the migration costs for the
> customer).
> 
> Though data are important and it is nice to collect them,
> I am still convinced that we help the developing countries better if
> we first help them to build up technology know-how of their own.
> They will be far more efficient with the data collection later on.
> 
> > Would it not be good if, each time an Intergraph or ESRI or Autodesk made 
> > one of these grand gestures regarding their software, that they also 
> > provided an equal level of grant-based support (both financial and 
> > technical) in the data collection and dissemination tasks!
> 
> No, the data are locked into a proprietary environment one way or
> another. This will be a disadvantage in the mid-term.
> 
> > Without good quality spatial data - who needs GIS software?  ;>)
> 
> Who can collect high quality data without a GIS? -> The decision
> on the techology comes first.
> 
> I am not saying that data collection has to be stopped now to first
> install Free Software. But the priorities for should be changed.
> 
> > p.s. In any case, most developing nations have a simple solution to overly 
> > priced software - they simply buy one copy only! "Software piracy in 
> > Vietnam is estimated to be 97% for business software" 
> > (http://www.ealaddin.com/hasp/) and estimates from the Federation Against 
> > Software Theft (FAST - http://www.fast.org.uk/) for the amount of business 
> > applications software that is pirated in developing nations ranges "from 
> > 50% (typically) to over 95%"!
> 
> This is just another argument to avoid proprietary software:
> As long as you are a poor country, illegal copies are tolerated.
> Once the country gets richer, you can collect as much money as you like.
> Also, it is possible to blackmail a country (think of a government
> using a proprietary oerpating system in all its authorities with
> only a few licenses - what position do they have for economic
> negotiations? China is a nice recent example, further will follow).
> 
> There are many other arguments why developing countries
> should avoid proprietary software to fill a book or a conference...