Slashdot Mirror

If you want to see open web mapping standards in action (and maps of the whole world) try here or look at some of NASA's data here

Ian

Client demands Linux

No commercial OS alternative

Even if we have the chance to use it, the requirement is that we do not actually deliver Linux with our code (the customer must obtain the Linux OS themselves and provide a clause to protect the project should actual IP issues arise). Again, this is a legal department requirement. Guess this will provide some additional level of protection.

No one seriously belives that there are "legal issues surrounding IP claims in Linux"

As much as users would like to believe no IP issues exist, in the legal world this issue hasn't been completely resolved. We all await a final decision on this.

Just an FYI, the company I work for does not have a problem with Open Source or Open Standards. In fact we support both (my group has some level of involvement in the Open Geospatial Consortium actually).

ESRI hasn't truly ported all of ArcGIS to *NIX platforms and I'm not sure that they will. ArcGIS uses COM for it's ArcObjects (which is what ArcGIS and several of the other offerings from ESRI are built on) and for ArcIMS they have included a COM application for use on UNIX systems to provide limited ArcGIS support (sort of like their "ArcMap Server for ArcIMS" that was only available for ArcIMS 4.x on MS Win32 platforms).

I have been using ESRI products since ArcInfo 6.1 (used on DEC Ultrix back in early 90's) and ESRI seems to pretty much own the GIS market. At my current job, I build/integrate various applications for geospatial/imagery and automated analysis and find that my customers really don't look for anything but the COTS applications with some small amount of "wrapper" code. I'd like to see Open Source GIS suceed, but don't know if it will in the corporate world where several thousand dollars isn't much compared to the time/money spent on working through an Open Source product. Eventually I see Open Source GIS working, but probably not in the next couple of years.

As for ArcGIS, I use it, but find that it runs way too slow. Give me an Arc prompt and I'll be much happier. ArcGIS consumes a good bit of memory and takes way to long to perform operations that were done much faster on the older ArcInfo software on much less of a system (CPU/memory).

Maybe the Open Source products will mature faster if open data formats are fully supported. Unfortunately ESRI only seems to do "lip service" to supporting things like Open GIS Consortium (OGC).

BTW, messed around with Ionic Software's RedSpider and it seems to do a decent job at serving WFS data, but it has a hefty price tag ($50K +). Can anyone recommend a good Open Source WFS/WMS server that can read either Oracle Spatial or ESRI ArcSDE data sources? I've looked at GeoTools and they don't seem to be "there" yet.

You should consider yourself lucky. Most of the information you need is already avaiable to you, in the public domain. As several other has mentioned, the Tiger-files will provide loads of information and you should really have a look at those (to convert them into your own format). It took a few weeks to do it, but the task is far from impossible.

The other question asked is however much more important, what about completly open maps in a free for all use setting? As i mentioned, this is the case for the US, but quite far from the truth for some other countries in the world :| If you're going to let other people use your data, please provide them in an open and accesible format, like WFS. Have a look at GeoServer and PostGIS (for PostgreSQL) or do as we do, store everything as GML - an open standard presented by OpenGIS (which also stands behind WFS and WMS). Take a look at their website which features quite a few important standards and other resources.

Making data available as WMS or WFS allows other people to seamlessly integrate them into their own applications. Seeing an application just importing more and more information thats available by WMS is simply amazing. The norwegian rescue service uses an internal WMS-server for all their mapping data, which provides information about currents, weather, available ships in some parts of the world that supports the system and loads of other information. This comes from several different sources and are integrated into the application on the last step. All the seperate units are responsible for their own mapping data and can upgrade and improve their data at any time without any interaction from the end user.

We export information by WFS, although probably not very interesting for your use, it demonstrates the possibilities. You may browse our repository at OneMap by using our SVG client.

GIS is a new-ish field, still developping very fast. A lot of tools are fairly mature, but the prices are still high, interoperability is getting good but there aren't many mature commodity components.

The major industry effort towards interop seems to be OpenGIS.

Some open source GIS stuff that looks promising to me are Mapserver and OpenMap.

I found the learning curve too much at this point, and many of the OSS solutions didn't work straight out of the box. Proprietary solutions are so expensive that they made playing around impossible.

What's more, getting data was difficult. Your city should be able to share its digitized maps. Here in Canada, my city was reluctant to share them, as some are copyright to ESRI (imagine your city co-owning its information with a foreign company!!!). What I found out however is that there isn't any copyright if you take the paper maps they publish and digitize it yourself. Time consuming, I know :(

There are a lot of useful hacks that I wanted to do with geographic data, but I shelved those plans for now. Hopefully in a year or so we will have better tools and cheaper data. If you manage to help us get there, thanks in advance :)

Sorry for preaching about GIS. My main argument is that if you want to get your data to the masses, you publish it in a format that the masses can read. GIS data isn't generally the type of data that most people would be interested and the software development reflects that by being a higher cost. GIS software, namely the main stream market leaders, is quite expensive due to the fact that development cost is spread over a smaller number of sales.

ESRI certainly has it's foot in the door of most agencies that are interested in producing geographic datasets, but they aren't the only ones. Intergraph, another major GIS vendor (I forget which position they are in for sales/installations), competes hard with ESRI to sell their software and also practices the same strategies of encouraging their format. Intergraphs GeoMedia software (maybe it was a different name before) started out on their own proprietery hardware called Clipper workstations which NIMA (NGA now) was locked into at one point. The parts all came from Intergraph and they controlled the pricing and had all sorts of political push from the representatives in Alabama .

Until recently (say the last 5 years or so), the computing power to work with these national datasets was limited to a small community of agencies/companies. Consider the data format that an agency chooses to be akin to distributing a document as either MS Word format or Adobe PDF. If an agency provides the data, it has to be in some format and they typically choose one that reaches the broadest audience. That being said, there are organizations like Open GIS Consortium that are working to create open standards so that the smaller GIS vendors and the not so traditional GIS vendors can have access to data once reserved for the more powerful systems. A company called Ionic produces an application called Red Spider (NOTE: This software is very very expensive. We evaluated it and found while it impliments most Open GIS Consortium "OGC" specifications, we could never pass the cost on to our customers.) that allows for "Web Feature Service" or "WFS". An open source project that is attempting to impliment these specifications, and certainly not the only one, is GeoTools, but they are slow in development and not mature enough at this point for commercial deployments. WFS allows applications to request geographic data similar to other web service applications. Part of the problem with these open standards though is getting people to buy into them. Obviously vendors like ESRI, Intergraph, Microstation, etc... have lesser interest in promoting the open standards over their own product standards, but they are starting to do it.

As far as restrictions on the datasets, I'm not entirely sure I follow you on that. The datasets are generally available for free (sometimes export controlled, but still free) and free viewers do exist. Granted that the free viewers are not always the best product for "high end" analysis, but they do at minimum allow for data viewing. One such tool from ESRI is called ArcExplorer (This link provides the location to download the viewer and an Interoperatability Extension) that is available for various platforms (MS Win32, Linux, UNIX, Mac) as it's Java based.

I know there are other open source GIS analysis applications, but I haven't used them as my employers have had the ability to purchase the high end commercial software. If I recall correctly, a copy of the ArcView software was less than $1000 a couple years ago and the software was available on UNIX (Sun, SGI, and a few others) and MS Win32 systems (sorry, no Linux version that I'm aware of).

I think I jumped around a bit on my above ramble so please forgive. I am coding/compiling and writing in between things. I guess I'm not sure if your disappointed a

Ian

Centre for Computational Geography, Uni of Leeds

Second, data providers should be publishing their data using OpenGIS standards such as Web Map Services and Web Feature Services, so I can use any OGC compliant interface (or implement my own).

My favourite online mapping software is Mapserver because it's open source, and compares really well with any of the commercial offerings.

Of course, good software counts for squat unless you have good data behind it. Good geocoded address/driving data takes time and money to compile.

Xix.

• Sensor Webs at Nasa
• Sensor Web Whitepaper
• PC Magazine article

The data could be made available to the public through a Web Map Server following the Open GIS Consortium specs OpenGIS.org.

This data could then be incorporated into your own personal Web enabled apps utilizing an open source product like U of MN Mapserver

<element x="0" y="0">
<component name="red" value="10" /> ...

Yep, the standard XML-bloat joke. Here is the serious solution: binary encoding. I have done some testing by defining a simple demonstration XML image format as:

<XmlDemoImage version="1.1.0">
<Header>
<Width>x</Width>
<Height>y</Height>
<SampleType>byte</SampleType>
</Header>
<Scanline row="i"> <!-- optional attr: filter="diff" -->
<RgbSamples>r g b r g b ...</RgbSamples>
</Scanline> ...
</XmlDemoImage>

If we can avoid the bizarre and hugely self-defeating but all-too-common urge to way-overstructure the pixel representation and use raw binary encoding especially for the dense arrays of numbers, the representation and performance is essentially equivalent to that of PNG format itself (though for some images, BZIP2 compression is significantly better). Here is a study of the issue. On an Athlon-XP1800+ Linux box, I get a raw (Binary)XML reading speed of 188 MB/sec for an uncompressed image. W3C is holding a workshop on binary XML encoding in September, so it may finally be prepared to humour such radical efficiency with XML.

I still don't think so. XML is really verbose, and there are plenty of advantages (and disadvantages) to binary files.

Have both.

Thanks - very interesting reply. I am interested because I am a transportation consultant in Oakland, with a focus on ITS and open standards. Cool that California DOT is using open source! We have done work for them in the past where they have been less enthusiastic.

If you were curious about the details of the remote traffic and weather monitoring stations
I know a little about this, I was more curious about any sort of web services exposure. For example, will you expose traffic information using something like SOAP or (even better) using something from the OpenGIS Consoritum? Not a big deal, just wondering what kind of cool services I can build on top for my clients. Thanks for the reply, btw.

Check out BXML for a binary encoding of XML to efficiently carry scientific/array-type data. Feedback appreciated.

It would make more sense to set a standard that would let people use DNS to deliver physical mail. That way mail routing could be altered on-the-fly depending on what where the destination was.

And don't get me started on Map Point. Thre's a perfectly good set of *extensible* standards for storing, querying and retrieving spatial attribute data. Map Point looks like a half-arsed set of proprietary spatial datatypes that they acquired a couple of years back in order to duplicate these services (start by looking at www.opengis.org)

Xix.

• Microsoft Terraserver.com is one of the big ones, selling images from lots of satellite sources. Originally a 1998 joint venture with MS, USGS and Compaq. Free lower-res stuff, subscription medium-res, high-res pictures for sale.
  
• SRI Digital Earth -
  Talk
  - DARPA project, some good stuff.
  
• LivingEarth.com and EarthImaging.com - more hi-res maps.
  
• Fourmilab.to Earth Viewer also does satellites, stars, etc. Slightly overworked due to Iraq conflict so using lower-resolution pictures.
  
• OpenGIS.org - Standards for geo-enabling web and other apps.
  

Now indeed, various groups/individuals have been attempting to implement these OpenGIS standards as best they can given what is available. You might have a look at:
geotools
OGR Simple Feature Library

While not OpenSource, the following is an interesting implementation of an OpenGIS compatible map server:
DEMIS World Map Server

(This OGC Web Map Server spec can be found at: 00-028.pdf)

Opengis.org is the best organizational reference for open GIS standards. They have an international consortium of business and government agencies behind them. They have been around since 1993 and have developed several standards for developing a true open framework for GIS delivery. In fact, GIS is one of the rare applications that demands a very open approach since having geographic data is only useful as it relates to other geographic data.

Opengis.org has done a good job of specing out systems that are truly interoperable because they achieve GIS nirvana: seperating content from visualization. Reading GIS content from multiple servers and displaying it through a single user interface is the heart of open GIS. Amazingly, no major commercial vendors (ESRI, Bentley) are aggressively pursuing this vision. IMHO, this is an opportunity for the open source community to make a mark on a major emerging industry! If you are interested in working on developing an open source version of the server spec that Opengis has released, please contact me!

--

I've been working on the Open GIS spec for web mapping for a little less than a year now. We've come up with a simple spec that lets you construct a URL with a map request in it. The spec describes the format of the URL and how map servers should behave when they receive one of these.

If you build a campus map server that conforms to this spec, then later on, you can build HTML pages that can request campus maps superimposed on satellite photos or other neat data.

The first place to look is at the spec. Revision 1.0 is due out "any day now" so don't get hung up on the details. Then look at what some people are doing with the spec.

Here's a sample HTML client that talks to a Web Map Server which itself then knows about a dozen or so other Web Map Servers. (What we call a "cascading map server").

Take out the nospam from my email if you want to ask more questions... (By the way, I'm also a big OpenMap fan - I open-sourced it while at BBN...)

I've been working on the Open GIS spec for web mapping for a little less than a year now. We've come up with a simple spec that lets you construct a URL with a map request in it. The spec describes the format of the URL and how map servers should behave when they receive one of these.

If you build a campus map server that conforms to this spec, then later on, you can build HTML pages that can request campus maps superimposed on satellite photos or other neat data.

The first place to look is at the spec. Revision 1.0 is due out "any day now" so don't get hung up on the details. Then look at what some people are doing with the spec.

Here's a sample HTML client that talks to a Web Map Server which itself then knows about a dozen or so other Web Map Servers. (What we call a "cascading map server").

Take out the nospam from my email if you want to ask more questions... (By the way, I'm also a big OpenMap fan - I open-sourced it while at BBN...)

I've been working on the Open GIS spec for web mapping for a little less than a year now. We've come up with a simple spec that lets you construct a URL with a map request in it. The spec describes the format of the URL and how map servers should behave when they receive one of these.

If you build a campus map server that conforms to this spec, then later on, you can build HTML pages that can request campus maps superimposed on satellite photos or other neat data.

The first place to look is at the spec. Revision 1.0 is due out "any day now" so don't get hung up on the details. Then look at what some people are doing with the spec.

Here's a sample HTML client that talks to a Web Map Server which itself then knows about a dozen or so other Web Map Servers. (What we call a "cascading map server").

Take out the nospam from my email if you want to ask more questions... (By the way, I'm also a big OpenMap fan - I open-sourced it while at BBN...)