Slashdot Mirror
Open Source, GIS and Data Visualization?
Mubarmij asks: "A lot of people, including the ex-Vice President, think that Terrain Visualization and the Georeferencing of all kinds of data is the next big thing. Given the broad applications (sims, education, games, GIS, virtual
tourism, etc) that can be derived from such technology, I would tend to agree that if this is not actually the NBG(tm), then at least it is very close. Like the internet, this technology has taken its time in obtaining it's current level of sophistication. However,
there is huge potential here that has yet to be tapped, despite the fact that it currently fills a huge niche market. I had once read that NASA spends more than 70%
of its resources on space imaging and visualization-related activities (unfortunately I link to the article that mentions this, but one should remember that the major goal of all space satellites is to take multihued pictures of Earth and other planets, and you will see that this is not an exaggeration), which is quite a
lot of money." Open Source has provided several frameworks for GIS from which a "killer app" may spring from. Read more on the various Open Source projects on GIS, and feel free to share your thoughts on where this technology may head in the future.
"There are quite a few web sites, commercial and non commercial that tend to this technology. However, it seems like the early nineties, where people are just starting to get aware of the internet, but are still awaiting for the killer app to make this thing fly.
There are two open source projects I am aware of that deal with this area. The first, VTP, is a real open source project attempting to create a real terrain visualizer. The second, OpenSkies, is not really open source yet (despite its claim)... but it is interesting in that it allows networked people to fly or drive through virtual worlds that are reality based.
Here are a few other questions:
- Do you think that this technology will remain a niche market (albeit a big one)? If so, is this likely to occur?
- Are you aware of any open source projects other than the two mentioned above that deal with this area?"
The really sad thing is that I worked for ESRI's big competition (GeoVision) in the 1987-1992 time-frame. Back in 1991-2 I got interested in Linux, and I actually considered stealing a copy of the source code from work to try and port it to Linux (which I would have given back to the company). It wouldn't have been a big deal, since it ran on SunOS, Ultrix and AIX. The only problem at that time was the lack of a decent SQL database for Linux. Mostly our stuff ran on Oracle, but it also ran on a few others. The other major problem was that the only way I could transfer all the source code for a huge GIS application from work to my home PC was on 3.5" floppies. That would have been more painful that installing SLS 1.03 from 5.25" floppies, which I had also done.
GeoVision's product is still around, after going through a series of owners. It's still a damn good GIS. And it would probably be dead easy to port it to Linux now, since Oracle is on Linux. Heck, I think even Oracle MultiDimension, a product I worked on at Oracle, is available for Linux.
The next Cmdr Taco duplicate will be ready soon, but subscribers can beat the rush and see it early!
Human beings are pattern recognizers. Our hardware is wired best to look/listen/feel/taste and match that stimulus to some other previously established pattern. We put a lot of effort into data visualization because it is easier for us to look at a picture and create a pattern than to look at raw data and create a pattern. Even the most difficult mathematical phenomena (noncommutative groups) make better sense when you give someone a concrete or highly visual example (rubik's cube, arrangements of books, etc.)
Sometimes the most brilliant of mathematical work comes from building a simple model that greatly simplifies the work (think Feynmann diagrams).
This can even apply to theology: when God intersects in our lives, we try to create a pattern to explain what God is, even though God is unfathomable. And thus we end up with many people who end up believing in Jesus or Allah or Nirvana, and then some people who end up believing in the UFO behind the comet. This also uncovers the problem with pattern matching: sometimes the patterns we find are dead fucking wrong (Aristotle). Fortunately we also have scientific method and reason to weed out self-contradicting, nonsensical, or otherwise false patterns.
ObJectBridge (GPL'd Java ODMG) needs volunteers.
Finding God in a Dog
As for the field of displaying quantitative information, the recommended books are Tufte . It is actually quite hard to create intuitively understood data visualisations because our eye-brain can only measure simple things like intensity, distance, etc. That's why things like pie charts where the angle is directly proportional to the propoertion works. All the other data visualisation techniques (parallel coordinates, tensors, etc) require a fair amount of training and patterning before you can pick up the meaning. A geologist (or related discpline) would be able to look at a contour map and be able to "see" in eir mind's eye the slope and elevation. Lesser mortals would probably require a pan of a 3D VRML model and even then have difficulty in recalling specific features. Adding extra layers or texture maps might be aesthetically pleasing (cough*QUAKE*cough) but doesn't really add any extra information.
LL
It's mainly for the folks who need a 2D, Java-based solution. It's pretty darn useful; lots of great projection math utilities, a nicely structured architecture, support for a wide variety of GIS products out there, and has transparent support for delivering imagery over a network. Plus the platform of choice for primary developers (BBN) is Linux. :)
http://openmap.bbn.com
everyone remember the challenger explosion? anyone remember that a couple of NASA scientists had done a bunch of research on o-rings and had determined that under specific conditions, that they wouldn't work as they were supposed to? but the challenger launch wasn't stopped, because the research was presented with such poor visualization that the managers of the projects (who didn't really understand the science) couldn't make informed decisions. part of visualization is helping really, really smart people show off their ideas to the rest of us.
I am the king... of No Pants! www.penny-arcade.com
I have tried for four years to get the GIS industry leader Environmental Systems Research Institute (ESRI) to port its products to Linux. They have expressed interest, but so far, have ported zero product. The problem is that they have not only leadership, but actual dominance in the industry You can send email to jdangermond@esri.com. Jack Dangermond is the president of the company, and a reasonable man. He has stated in the past that product ports are driven by user demand. Please be polite, ask for the specific product, and refer him to this slashdot site. I think that if he sees that the userbase is considering jumping to Open source, it might help his decision. Thanks Greg
The reason, is simple.
Most beginning projects focus on the purchase of GIS software, like ESRI's ArcInfo, the purchase of sexy machines to run it, and perhaps some initial staffing to set it up. They almost uniformly massively underestimate the amount of time and money required to actually get a functional landbase into the system, attach data to it, and ensure the necessary level of accuracy of the basemaps.
"It'll be easy, we have DB2 running on an as/400 and lots of maps" - should qualify as one of the most expensive single instances of "famous last words" in computer history.
In reality, very few maps are accurately digitized. Most GIS systems get sold on the premise of scanning and vectorizing dead-tree bluelines, or other formats - from paper. the problem is, that geographic maps are finicky about scale, and the maps you *thought* were accurate inevitably need to be rectified to whatever coordinate system you chose for the system. Its a tedious and enormously time-consuming process that simply doesn't lend itself to any sort of automatic processing. Then, there's the problem that most maps simply aren't accurate to begin with, and you see the beginnings of a problem. "I just bought half a million dollars of GIS software and equipment, and now I discover it's going to cost me three or four million to get my maps in shape to actually be useful".
Picture your average City government. Municipalities are where government really is. About 90% of the civil services provided to you by the sum of all government influence on you, comes from the local level. Cities have huge quantities of maps for things like building plats, subdivision maps, building blueprints, deed records, thoroughfare maps, land-use plans, etc etc etc. And that, coincidentally is where GIS appears to be the most useful. managing all that data is hard for cities - and GIS certainly looks easy in the presentations companies and consultants present - just point and click! What could be easier?
But the numbers of abandoned systems started by cities who bought into these sales pitches are huge - on the order of billions of dollars worth.
GIS companies and consultants know this - yet never warn purchasers, that in most cases, the cost of the software and systems amounts to only a tiny fraction of the actual costs of developing a working GIS.
So when you hear all the gushing success stories, and gee-whiz ideas presented as though GIS was a wonder cure for all sorts of problems, try to bear in mind - GIS works best with single-source maps, and data which can easily be applied to pre-existing points. If you hear the words "scan" or "address-range" - you'll know a bullshit artist is at work. Because for every working system involving simple maps with pre-rectified sources, there are probably 15 that were simply abandoned because of enormous unexpected costs and (salesman provoked) unrealistic expectations. A lot of money has been pissed away on GIS over the years. Real computer professionals know the golden rule - flashy graphics doesn't equal useful purpose.
The other cool thing about it is the protocols and data formats it uses to transfer information in both it's 2D and 3D mode are publically available. The idea is in the same way Palms are popular because anyone can develop for them, that's the goal with this product.
In fact, the 3D client is Open Source!
I hope this might be some where close to the kind of tool you're looking for.
antarctican at trams dot ca