Open Source Geographic Information Systems

RGillig writes "The second MapServer Users Meeting and the first ever Open Source GIS Conference was held on June 9th to 11th in Ottawa, Ontario, Canada. The initial response from the Open Source GIS community is that the conference was a huge success. It was great to have people from private, government, academia, and communities all together discussing how Open Source GIS applies to their needs. Here is a presentation given by Paul Ramsey, Director, Refractions Research Inc. that outlines the current state-of-the-art for Open Source GIS, and includes links and information about all of the current software packages/efforts, etc."

Comments

Where is Ottawa, Ontario, Canada?

[toetagger1]

Isn't it ironic that they have to specify that Ottowa and Ontario is in Canada, when the whole article talks about maps?

Re:Where is Ottawa, Ontario, Canada?

[xs650]

They realized that many of their geographically challanged friends from south of the border would read the report.

Re:Where is Ottawa, Ontario, Canada?

[0racle]

I know, its so strange, its not like there's any other Ottawa or Ontario anywhere else in the world. Yep, only ones are in Canada, and while were at it only Americans don't know geography. Any other cliches you want to throw in?

Re:Where is Ottawa, Ontario, Canada?

[RollingThunder]

It does happen to be our nation's capital. That would tend to make it the default assumption for large meetings like this, rather than smaller places like you mention.

If I say "there's going to be a major convention in London", I would assume London, England - not London, Ontario, Canada - and expect others to assume the same.

Re:Where is Ottawa, Ontario, Canada?

[0racle]

Never assume anything. A large portion of the Canadian population doesn't know that Ottawa is the capital, Toronto is the first choice of many. Canadians don' know any more about geography then anyone else.

Re:Where is Ottawa, Ontario, Canada?

[Jardine]

A large portion of the Canadian population doesn't know that Ottawa is the capital, Toronto is the first choice of many.

You seem to have misspelled American. Do you have a source for this statistic?

Re:Where is Ottawa, Ontario, Canada?

[whiteranger99x]

Well there IS a City of Ontario in California, USA

Granted while it a little out of scope, it indeed proves that there IS another Ontario anywhere else in the world :)

Re:Where is Ottawa, Ontario, Canada?

[xs650]

" True though, try asking some random person how many states are in Canada some time."

The only one I can think of off hand is North Dakota.

I tried to go to that

[Anonymous Coward]

Couldn't find the damn place though.

open source GIS

[Anonymous Coward]

So when is GIS going to be tied into the internet so when I search for a pizza joint the first result won't be a place that's 300 miles away?

Re:open source GIS

[LostCluster]

I'd suggest sailing your ship a little closer to the coastline first... there aren't very many places that deliver pizza to ships in international waters 285 miles away from the nearest land mass. :)

It happened when?

[rikkards]

I live in Ottawa and never heard about it. Hmm. Maybe it was due to the fact that the two big summer events that anyone talked about here were the Hope beach volleyball tournament (today) and Bluesfest (which started yesterday).

opensource GIS predates Linux...

[prof_peabody]

Many of you may have forgotten that GMT (generic mapping tool) is open source and predates linux. I'm glad to see more opensource work in the GIS field, as many companies charge bundles of cash for very basic GIS software.

This is good stuff

[goatstuffer]

Outside of the end-user type applications (ESRI's ArcGIS and co.), open source in GIS is quite widespread.

Refractions Research maintains the PostGIS module for PostgreSQL, and while it is not yet complete (fix the ACROSS function guys!) it certainly makes the wrangling of data much easier as it implements the OpenGIS SQL specification.

Compare this to the old days of a dozen different formats which weren't convertable, it's much nicer with GML (Geographic Markup Language) and standard representations of geographic features made possible by the find folks involved in the OpenGIS consortium.

Props to the team at the University of Minnesota for MapServer, it's made my life a whole lot easier.

Re:This is good stuff

[temojen]

Indeed. I was glad to read of JUMP in this report because I was looking for something along the lines of ARCView for occasional use and had been very frustrated with GRASS. GRASS may be extremely powerful and flexible for geographers etc, but for occasional analysis (by non-experts) it really sucks. JUMP looks to be just the ticket.

I guess I'll know a little while once the download completes.

Re:This is good stuff

[Jon_Aquino]

I'm on the JUMP development team and I hope it meets your needs. It's a good program for editing 10MB shapefiles. It can also edit GML, though not as easily. And it has a simple Java plugin system, so you can make it understand any data format (or database) (or do anything for that matter).

Feel free to contact me or to sign up on the mailing list for the JUMP Unified Mapping Platform.

Texas Mesonet

[Anonymous Coward]

I did a lot of work with MapServer and GIS data at Texas A&M for a part-time job I had my last semester (this has been close to 2 years ago now). Check out the Texas Mesonet project at:

http://mesonet.tamu.edu/

Click on Current Weather to see the MapServer-based map I helped create initially. It's all built with open-source software and (I think) freely available data from the national weather service. It's amazing how much data you get, and how easily it can be handled by one little machine in a windowless office somewhere (until it's slashdotted of course).

doc file?

[shaitand]

Anyone else have a clue why information about an open source anything would be in a proprietary MS format?

Re:doc file?

[shaitand]

That's hardly the point is it?

It's not exactly much of an endorsement for OSS if you use proprietary formats to distribute that endorsement.

I really wish people would stop pretending that simply because openoffice reads docs is some valid justification for using the format. For one, there is nothing you can put in a doc you can't put in an open format, and most of what is put in a doc should be put in an rtf or txt file.

For another, there is no guarantee MS won't change the doc format tomorrow, which is the entire point of NOT using their formats.

Programmers' tools, not finished applications

[isdnip]

The "briefing" has a good collection of pointers to open-source applications out there. But as a fan of the commercial Windows GIS product MapInfo, I am frustrated by the lack of an open source alternative, and by the lack of comparable tools for Linux. GRASS is pretty powerful, but it's not something anybody can just start using; it's more like something a Unix GIS professional (difficult but powerful systems like ESRIs) would find interesting.

This note from the briefing is most telling:

Note: The saturated commercial market for cartography tools, the high level of effort to achieve a usable tools, and the appeal of other cutting edge projects have combined to deter any active development on user-friendly paper map production tools. As with the OpenOffice experience in Linux, it would probably require a dedicated multi-year funded project to produce a core product with sufficient technical mass that an open source community could reasonably continue with enhancements and support.

In other words, don't expect to find a complete open source end-user application within your lifetime.

This is, alas, common in the open source world. Everybody does their own toolkit that does 90% of what other toolkits do, adds 10% of its own, and assumes that the user is a person who gets their jollies from writing code, not actually using the application with production data.

Re:Programmers' tools, not finished applications

[Frank Warmerdam]

> In other words, don't expect to find a complete
> open source end-user application within your
> lifetime.

The comment you quoted addresses the specific topic of cartographic map generation suitable for printing. I don't see any reason that several of the existing projects can't include respectible map production suitable for most GIS end users.

Furthermore, as noted, a serious cartographic production system could be implemented within a couple of years given an appropriate project to drive it.

> This is, alas, common in the open source world.
> Everybody does their own toolkit that does 90%
> of what other toolkits do, adds 10% of its own,
> and assumes that the user is a person who gets
> their jollies from writing code, not actually
> using the application with production data.

Frankly, the report indicates that there is a great deal of sharing of supporting toolkits between the end user applications of various kinds. I think the open source world is much less to duplication of effort than the proprietary software world.

Also, many of the required applications do not require the end user to write code to do work.

There is still some way to go before any of the software packages is across-the-board competative with software offerings from ESRI or MapInfo but I would like to think that for end-user applications for much typical GIS work is coming together now. And many specific tasks are already filled better by open source tools in in this space, than by commercial tools.

In short, I feel your claim these are programmers tools, and not finished applications is unfair though I will conceed that none of these applications (with the possible exception of GRASS which has some ease of use issues) has as large a feature set as the major proprietary packages.

Re:Programmers' tools, not finished applications

[geodude]

There are a number of recent efforts in the exploratory spatial data analysis real that attempt to deliver on the "end user" side of things, for example:

Geoda http://sal.agecon.uiuc.edu/geoda_main.php

STARS http://stars-py.sf.net

Choro http://choroware.sourceforge.net/

Re:Programmers' tools, not finished applications

[SendBot]

I use MapInfo at work, and was toying with a custom alternative using SVG output of the image_gis php module. More info here

I didn't get very far as the documentation is pretty light, and I have a hard time coming up with info in the Arcinfo/E00 format or finding a decent converter.

Re:Programmers' tools, not finished applications

[KjetilK]

GRASS is pretty powerful, but it's not something anybody can just start using; it's more like something a Unix GIS professional (difficult but powerful systems like ESRIs) would find interesting.

That's very interesting! I was wondering if you could give me some advice...?

This is the situatation: I'm looking at GIS now, as I need to expand my skills, and only solutions running on Linux will come under consideration. Furthermore, I wouldn't trust systems where I can't inspect the source code. It doesn't need to be free as in speech, but the source code must be available.

I've looked at GRASS, since it is in Debian. It segfaulted on me when I tried to load a data set, so I didn't get very far. It did indeed look rather hard to use, but since I am a long time UNIX user, and can do some hacking myself, perhaps it is for me anyway...? I'm also a long-time R user (I love that system), and the two are supposed to work well together.

So, what you're saying is that GRASS is a powerful system, but has a steep learning curve?

That's quite OK by me... But does it flatten some time? That is, is it designed so that when you've grokked the fundamentals, you can pretty much do anything?

I think what I'll do the most is to create topographic maps from DEMs. Then, I may do some tracings of LANDSAT or ASTER data, to add some rivers, glaciers and stuff like that. How hard would this be?

Its good, but not the complete picture

[PierceLabs]

What we need are good royalty and free-use datasets that allow open source products to actually be able to do high resolution GIS queries. Without a large volume of free data, having an open source GIS system isn't enough.

open source?

[IGnatius T Foobar]

Name : 2004-05-OSS-Briefing.doc

These "open source GIS" people need to learn a few things about "open source software." Presentation in Microsoft Word format? Faux pas!!

Refractions Research = excellend support

[Anonymous Coward]

Just want to say that Refractions Research's postgis mailinglist is one of the best customer support experiences I've ever had. A prototype of one of our future products (crime mapping software) is based on PostGIS, and 4am the night before a customer demo we were having some problems (postgres optimizer on geom indexes).

By 4:30 AM we had exchanged about 3 emails each way, fixed all the problems and had a great demo. If we land the client, we're hiring them.

Here's Hoping

[Nilmat]

Back in the summer of 2001 I used GRASS pretty extensively. At the time, it could do a lot of the same stuff as ArcView and ArcGIS but was vastly clunkier in doing it. Think Gimp vs. Photoshop a few years ago. I'm glad to see that open source GIS lives on, since a workable alternative to ArcGIS is absolutely essential for those of us in academia. In fact, I've given up on ArcGIS and still use ArcView because I can't stand the damn thing. It also doesn't help that you can't run ArcGIS under anything OS but Windows, since its all written in VB. I've even tried to run ArcGIS under Windows via VMWare, but it doesn't recognize the necessary hardware key. Enough with rant there, but in any case I guess I'm just hoping that one of these open source alternatives will be viable in the near future.

This would be a good thing for WiGLE

[ONU CS Geek]

I use and upload information into WiGLE (wigle.net), and having information like this would do wonders in having accuracy in mapping and plotting. There ahve been times where I've plotted information, but the information from Tiger isn't up to date, so my plots don't look like they're on roads.

Now, if we could only work on GPS accuracy. Sure, 21 feet is 21 feet, but, still...I'd love to be able to wardrive and know exactly where something is at. (Yes, for the subtle, I know that 21 feet doesn't make much of a difference with a Wi-Fi point, but, being able to accurately identify where a point is would be nice. Instead of knowing where on Randall Road something is, it'd be the bomb if we could pick up something like 4033 Randall Road from the GPS Coordinates.)

Maybe I'm just dreaming, or had one too many to drink on a Saturday night.

I wouldn't trust an open source mapping system..

[Anonymous Coward]

because I expect that where Redmond, WA is, the map shows a giant lake. :)

It's out there.

[sp0rk173]

As an undergrad researcher currently doin a heavily GIS-intensive project, i have to say the data is out there. In the US, the USGS provides multitudes of data for free, as does the EPA (the BASINS dataset is HUGE and completely free). Granted, it's hard as fuck to track down if you don't know someone who has already had to sift through the many, many websites out there that hold the data - but it's out there. What needs to be done, I think, is for the community to create some kind of central portal that makes it easy to find, and then download all of the data. THAT would be helpful.

We have plenty of 'free' data...

[jim_deane]

There is a /ton/ of 'free' GIS data available on the internet.

I say 'free' because in reality the US taxpayers have paid for it, but take a look at things like:

Kansas DASC,

Census Bureau TIGER data,

collection sites like Geo Community,

and an almost limitless number of other sites. Most states now have GIS sites of one form or another, with downloadable data.

Jim Deane

Re:We have plenty of 'free' data...

[geekopus]

I work for the Georgia (USA) GIS data clearinghouse. We have thousands of free datasets, and very reasonable pricing on downloadable imagery ($5 per USGS quarter-quad). Of course the imagery is kind of old (most recent is 1999 color infrared), but many people still find it useful. You do have to sign up, but believe me, it's not for any sinister purpose. There's only two people with direct access to the data; me and the guy in the office next to me. We don't do anything with it but collect aggregate statistics so the state of Georgia can determine how best to fund data collection by who is using it and how often it is being used.

The clearinghouse main page

Direct link to the data

There's also good imagery on the USGS site for free, but you have to use their viewer to view it (it's not downloadable...).

And as far as open-source GIS, has anyone here tried GeoTools? That's the most complete OSS GIS toolset I've seen.....

Data, not programs

[Anonymous Coward]

right now, what Canada needs is free access to high-quality current GIS data. The US has Tiger, we have nothing similar.
It's all controlled by municipalities. Toronto wants a small
fortune for copies of TAXPAYER paid-for data.

Re:Data, not programs

[Anonymous Coward]

Take a look at this, it has most of the canadian road network available for free (as in speech): http://www.geobase.ca

What about the rest of the world?

[buchanmilne]

Sure, the US has a lot of free GIS data, but maybe you've heard that there people who live outside the US? And, maybe they also prefer free software, open formats and more available data?

Re:What about the rest of the world?

[KjetilK]

There is at least the SRTM-3 data set. It is an excellent data set covering most of the landmass between 60 N and 58 S (which, unfortunately just barely includes me...). It has a spatial resolution of about 90 meters and an elevation resolution of about 15 meters.

It's in a simple binary matrix, easy enough to hack up something to import it whereever you want.

THERE'S NO DATA!!!

[Anonymous Coward]

For at least 90% of the world, there is no free GIS data at useful resolution. There's free data for most of the US, but not much for the rest of the world.

I'd say free data is the real issue, not free software.

Re:THERE'S NO DATA!!!

[kelk1]

You are so right. I think there are two reasons for that. First, the knowledge of positions has been and is somewhat still considered a military secret. Many mapping agencies (all?) come from a military background, so they are naturally reluctant to reveal the data. Second, they are making loads of green by selling the information to professionals who need them (e.g. surveyors).

Re:You don't have a clue

[SlamMan]

So, what, they were supposed to put it in an .sxi people without Open Office (read: almost everyone that has a computer, certainly most their audience) can't read it? Supporting open source doesn't mean everything you do needs to be open source. Because I work with apache, and advocate it over IIS, doesn't mean I'm going to chuck Dreamweaver and Photoshop, or not deal with .psd files.

Just because the format is open doesn't mean its the best solution. I'd say they should have made a pdf, but thats just me.

And today is?

[hswerdfe]

July 11th, it would be nice if someone would have told me about this ahead of time.. I live in ottawa and would have loved to atend

Not to burst anyone's bubble, but...

[paanta]

As much as I'd love to see ESRI relinquish its stranglehold on the end-user map-making world, I don't think I'll see a good, open source alternative for a _long_ time.

I've worked for one of the largest regional planning agencies in the country, for a ~100,000 person city, with planners and environmental types at at U of Michigan, and done a fair bit of GIS work on my own. ~95% of that work has been with ESRI products. Except for some specialized spatial statistics software, and equally specialized transportation modeling packages, ESRIs stuff is (sadly) hard to beat.

The (paying, non-researcher) end-user, a GIS lackey in a planning office somewhere, someone doing work for some environmental group or maybe someone doing marketing analysis, is not going to deal with the hassles that most open source packages involve. The most successful open-source end-user programs tend to be things with a _huge_ amount of interest in them. You know, web browsers, mail clients, desktop publishing, etc. GIS is still kind of a niche market. Maybe I'm totally off-base in assuming this, but my feeling is that ESRIs core customers are the big metropolitan planning organizations and those are _incredibly_ slow moving organizations for the most part. IMO, there has to be a lot of oomph behind a project before it gets polished enough that Joe Blow, Metropolitan Planner, is going to use it.

I love the idea of GRASS, but I don't see it ever out-doing ArcGIS. Open-source GIS needs to find a big, untapped market and branch out from there. I think what the open source GIS community needs to do is focus on a very stripped down package, as easy to use as a web browser, that lets the average person download TIGER line files from census, import ESRI shapefiles, add their own GPS data, with a big open source library of maps for people to play with. Leave out the analysis tools altogether, deal with things like map projection behind the scenes, and let people use GIS to plan gardens around their house, etc. Once you've got people using that, bloat the software from there, rather than slowly adding features to an already buggy, difficult to use package.

The other extreme of the spectrum is the high-end GIS work, where you've already got serious computer nerds working, and where there's always a market for a product that cedes some control back to the user, even if it is at the expense of some day-to-day usability. Thats where open source is already making inroads.

not a need for data, but a need for ACCESS to data

[goatbar]

Yes, so it doesn't take long to discover that there is a mountain of data available for free here in the US. The problem is GETTING the data. What a nightmare. The DataDepot is truely a hideous system. And ArcWeb (or what ever their web map server thing is) is totally frustrating to all but the most patient. Data comes from 10000 sources in 100's of formats and require a different way to get each one. Please don't make me separately click to download the 50 different files just to make a basemap of a new field area.

I've triend to make an effort to show how to do this, but it gets frustrating! You can see what I did here at my Visualization Classes. I used to be a Arc/Info hardcore user, but got so frustrated I gave up. It's easier for a programmer to write their own than deal with all the cruft in Arc. However, it's great for creating funny war stories.

if open source could trump esri

[enrico_suave]

If OSS GIS could get to the point where it can do one thing better than ESRI arc* products it would be a very good thing.

(not to say OSS GIS doesn't do certain things better than ESRI... let me explain)

If the OSS development community can build say, a viable online mapping platform that was open it would be huge!

I'm sick to death of the ESRI upgrade/maintance ladder/extortion to get product revisions that fix the bugs in the original release. I'm tired of the convoluted bandaid approach to online mapping.

I'd welcome a solid OSS solution any day, ideally beable to serve ESRI format .mxd files...

blah...

e.

Re:Where's the GPS?

[PipsqueakOnAP133]

http://www.sourceforge.net/projects/tmrs