|
THE
INTERVIEW
Five Questions for . . . Paul Ramsey
Paul Ramsey is the president of Refractions
Research, a company he founded in 1997. He serves as a
director, leads the development team, and serves as
systems architect. He’s also well-versed in open source
geospatial technology and open geospatial standards.
1. How do you explain the
concept of open source to potential users?
Open source software is software that is
distributed along with its source code. Software is
generally distributed as machine code, as a black box that
accepts certain inputs and generates outputs from them.
Open source opens up the box. Open source is like a
restaurant that not only serves up a tasty meal, but gives
you the recipes to take home and try yourself, maybe even
improve and change.
Whole treatises have been written on the general
concept of open source, probably the best is Eric
Raymond’s The Cathedral and the Bazaar which explores
how open source can be used as a tool of technical
community building, and how technical communities of
interest can in turn feed back and radically improve
source code through collaboration.
Open source enables collaboration between groups
and individuals that would ordinarily have no point of
contact. The UMN Mapserver, a Web mapping server (Figure
1), includes contributions funded by the government of
Canada, by NASA, and by the government of Brazil. It also
includes code donated by academic programmers, and
independent software developers. All these developers have
different goals, and are building radically different
systems, but the open source framework of Mapserver
provides them with a forum within which to collaborate.
By becoming part of an open source community, a new
user not only gets access to a very useful piece of
software, but a substantial peer group, who can share
experiences, and pool development effort to achieve common
goals.
2. What’s the role of open
source software in the geospatial market today? How’s it
“doing”?
As with open source in the larger technology
marketplace, it is generally invisible to the established
customers and practitioners. This is not surprising,
because organizations with established technology
environments and ways of doing things do not have a
compelling reason to look beyond their current
environments and suppliers.
It is around the edges of the market that open
source projects are making real inroads: with users new to
geospatial, who have no preconceived notions about tools
and technologies; and with established organizations
exploring relatively new application categories, like
OpenGIS Web services, where proprietary companies have
left some holes in their product lines. These new users
are taking up open source because it is freely available,
has no evaluation licenses to sign, and works.
As time goes on, and organizations reach the
end-of-life of their existing software and look at their
next generation software requirements, open source should
be considered a viable alternative, alongside other
products. We are already seeing this in the larger
marketplace—decision makers no longer look askance at
the open source Linux and Apache projects. In the
geospatial realm, mature open source products like
Mapserver and PostGIS [an open source product that
spatially enables open source PostgreSQL] (Figure 2) are
ready to be evaluated on their technical merits alongside
their proprietary equivalents—decision makers just need
to feel validated in doing so, and that takes time.
3. One of the concerns about
open source software in the geospatial community (and
others, I imagine) is how the companies that support it
make a living. It’s the time honored question: how do
you make money from something that’s free?
My company is an ESRI business partner, and one of
the ways we make money is by providing consulting and
development using the ESRI product line. The freeness or
non-freeness of a product is of little interest to a
consulting company. What is of interest is: does the
business equation of using this software make sense for my
client, and will my client pay me to implement a solution
using this software? So, the general answer is, you cannot
make money from something that is free, but you can make
money using something that is free.
However, your question was more pointed – you
want to know why a company like Refractions Research
specifically puts money and effort into products that
cannot directly return us any revenue. We do open source
development not to acquire revenue, but to acquire
clients, who will in turn provide us with revenue: either
clients who want to use PostGIS or another open source
product, and need our advice specifically about those
products, or clients who are simply impressed with what
PostGIS represents about our technical ability, and want
to use us for other geospatial consulting. Our work on
open source is a calling card which says “there are
spatial database experts over here, come ask them for
help!”
4. Your company, Refractions
Research, is behind the open source PostGIS. Do you think
its place in the “backend server” side of things
(Figure 3) makes it more likely be successful than desktop
GIS efforts like GRASS? Are we likely to have a widely
used open source desktop GIS anytime soon?
Open source projects that address the backend and
client/server geospatial computing problems have certainly
grown faster over the past few years than older projects
like GRASS that provide a monolithic desktop GIS. I think
this has a great deal to do with open market gaps, but it
is also indicative of the effort and complexity involved
in writing a complex desktop application like GIS
software.
GRASS was technically competitive with other GIS
products at the time of its inception (that was really the
point of GRASS, to do everything that proprietary products
could do and more) but that was many years before it was
open sourced. By the time it was open sourced, much of the
GRASS user community had moved on to other platforms, so
GRASS has been rebuilding a community from scratch, and it
has to be a community willing to put up with the
limitations of a platform that is getting close to
20-years-old.
Refractions Research is working right now on an
open source desktop GIS framework, that can efficiently
use the new OpenGIS Web services infrastructures, and also
integrate database servers and legacy file formats. The
project is called uDig (User-friendly Desktop Internet
GIS) and will have a 1.0 release in March 2005. uDig is
partially funded by the Canadian GeoInnovations program.
That said, I do not think there will be a
“widely used” open source desktop GIS any time soon.
It is possible that uDig will become widely used among
people who have never had access to a GIS before, but the
general GIS community has already committed time and
effort to learning certain tools, and it will take years
for a change over to some new ubiquitous tool to occur.
Whether the change over is to an open source suite, or to
something proprietary but completely unexpected, remains
to be seen.
5. IBM is leading traditional
commercial software vendors in turning over bits of its
source to the open source community. What does that choice
mean for commercial software players? For the open source
community?
IBM is quietly becoming a very aggressive open
source company. The company seems to have recognized that
the future of proprietary software is at the highest level
of innovation only, and that all other categories are
destined to eventually be standardized and commoditized.
So, the only people making money will be the extremely
innovative (and IBM continues to pour billions into
research and development) and the systems integrators (and
IBM Global Services continues to lead the pack as a
consulting organization).
For proprietary players, it means that one of their
own is turning on them, and actively seeking to
commoditize areas of software that might otherwise have
had a slightly longer run as proprietary domains. The IBM
donation of speech recognition code into the open source
domain is one aspect of this, as is the company’s
ongoing development of the Eclipse open source software
development environment.
For the open source community, IBM’s recognition
is a validation of sorts, and it helps in talking to some
decision makers, but it does not change the day to day
order of business. [For us that means] read the mailing
lists, help the users, think about suggestions, implement
the best, squash the bugs, talk with your peers about the
subject you love, and always make things better.
Back
|
