Technical Ramblings

This conference brought to light that I am one of the quintessential examples of a neogeographer: wordspy, for example, lists two examples for the term, one of which is:

Schmidt spends his time wandering around his hometown of Cambridge, Massachusetts, using his custom cell-phone software to unmask the ID numbers on each GSM cell tower he passes. Then he associates that tower ID with a GPS-defined location, and uploads it to his website.

When his electronic surveying is complete, Schmidt will have a system that can tell him where he is at all times — without GPS — by triangulating the signals from the newly mapped cell towers.

Calling himself a “neogeographer,” Schmidt is part of a generation of coders whose work is inspired by easily obtained map data, as well as the mashups made possible by Google Maps and Microsoft’s Virtual Earth.

—Annalee Newitz, “Where 2.0 Gives the World Meaning,” Wired News, June 16, 2006

A closer look at the article brings you to my pretty face above a screenshot of some work I did while doing cell tower mapping. Some people might recognize that the cell tower mapping in the picture is actually an image rasterized out of GRASS, an Open Source GIS tool which existed before the term “open source” did. GRASS has been around for 30 years, and is probably the single tool which is the clearest example of software from the “GIS experts” that the neogeographers are so often compared with.

So perhaps I am a neogeographer, but the image points out one thing only: even neogeographers can learn something from the GIS experts. This is a point which was made by Schuyler in his lightning talk at the FOSS4G conference, and has been made by me at other times in this journal of technical ramblings, but it’s a point that bears repeating.

I always said that I used the term neogeographer as a deragative term for myself, not one I expected to be used to praise me. After thinking about it more, I realized it’s not that it’s deragative — or it shouldn’t need to be. The term is describing the ability to take the things which other people agonize over, and make them fun. GIS work and neogeography are two ends of a spectrum: one dedicated to analysis and accuracy, the other dedicated to sharing stories, which oftentimes results in a loss of accuracy. That loss of accuracy is not always a bad thing: There are many things which don’t *need* to be accurate to the meter, or centimeter, level. Putting my photos on a map can tell a story — even without being 100% accurate to the meter.

Neogeography can be good — but many times, the neogeographers need to learn from the GIS experts before they reimplement the wheel, and give it squared off corners.

So, several times throughout the most recent conference, I found myself wishing that the 6×6 rule was a standard teaching item in more curriculum… until I realized that most of the people presenting have probably never taken any courses in presenting, so there’s no place in formal education it could fit. So, instead, I’ll share it with people here.

When crafting presentation slides, each slide should have no more than 6 lines, and each line should have no more than 6 words.

There are many reasons for this, but essentially it comes down to the fact that with more than 6×6, the text on the slide starts to to be the text of your talk. This leads to your audience reading your slides, ignoring you, and ending up bored.

The one exception to this I’ve found is in a case that I saw at this FOSS4G conference: the speaker spoke only French, but his slides were slightly more verbose… and in English. So even though he broke the 6×6 rule, it was effective: he was able to communicate his message in English with his slides, and his message more verbosely in French via the talk.

In general though, try following the 6×6 rule. You may find it helpful to prevent you from just reading your slides at conferences.

As I’ve so often found at conferences, I didn’t have enough time to write what I wanted, and now I’ve got all the time in the world to write it, and much less to write about, since nothing is happening.

The OpenLayers presentation was extremely well received. I did a survey of morning presenters, and OpenLayers was by far the best attended: We had about as many attendees as the other 6 sessions in the 8:30 slot put together. We had about 75 persons attending in total, and after our presentation, the room cleared out, which says to me that OpenLayers specifically was the draw for people. It’s good to know that the topic that we came here to present is a popular one.

Sadly, due to the early morning slot, no video of the presentation was made. This is especially unfortunate because there were so many people who were interested in the presentation, but unable to come. However, the slides for the presentation are available online (OpenOffice Impress), and the slides basically walk you through the presentation well. It was essentially a tutorial on doing a number of things with OpenLayers — adding WMS and commerial layers to a map, adding markers, adding popups, etc. I think we probably showed a lot of people just how easy it can be to create an OpenLayers map or application.

I want to thank anyone out there who attended the presentation, as well as MetaCarta for sending Schuyler and I over to this great conference.

Got a late start, so I missed the PostGIS case studies session I wanted to get to see. However, I’m currently at the QGIS as WMS Server talk. It is an interesting idea — essentially, they created a C++ CGI which calls out against the QGIS libraries.

It doesn’t really solve what I would like to see: a WMS implementation which is installable as a single binary app. You could open the data, style it, and once it was styled, you could serve that data — exactly as is — in any WMS client. The reason I want this is probably clear: this allows you to do OpenLayers development against a GIS application.

I think that Arc* are actually getting close to doing something like this, but that’s not a solution for me, since I’m looking primarily for something Open Source, and secondarily for free as in beer. Actually, even if it wasn’t free as in libre, I’d probably be happy with it.

One thing that’s hard to do right now is directly combine editing of GIS data and displaying it in a web mapping client. Certainly, one can see the uses that might lead to this: take an OpenLayers WMS setup, and you want to add data on top of it — perhaps you want to load a PostGIS datastore into QGIS, edit the points, and then immediately render the data into your webpage. If there were some GIS to provide a WMS against itself *directly* — not just via Apache or something else — you can skip a setup level of MapServer, or what have you.

MS4W is great, but it doesn’t include tools for creating geographic data. QGIS is great, but it doesn’t include tools for making it act as a WMS server. Maybe there are tools out there that do it — if so, maybe they’re Java, and we all know that I don’t touch that, but if it did this, I would.

I want to serve a WMS layer from a single GIS app running as a webserver on my local machine. What can do it? QGIS is apparently not the answer — you still need some CGI server to do the serving. So, what is it?

So, calling T-Mobile via Skype is an excercise in futility. After 7 calls, I did eventually get their voice prompting system to not immediately kick me out, and got through to a representative, but man, what a pain in the ass.

However, that does mean that international calling should be set up on my phone in the relatively near future. Not that I’ll use it for anything… but I can cellstumble with it 🙂

They also apologized for not setting it up when I called to ask about rates.

All in all, once you get to talk to a person, calling T-Mobile has been a very successful experience for me.

Data rates here are 1.5 cents/kb — that’s not nearly as bad as I would have expected, although it’s too expensive to make it generally useful. (If only the web weren’t so laden with heavy bandwidth crap. I blame DSL.)

This is my first trip over the Atlantic — I’m flying Northwest/KLM both ways, which is also a first for me. It was my first time in an Airbus 330 — I’ve almost never flown something with three sets of seats, other than on the round-trip to Washington DC, and that was 8 years ago. The flight east was uneventful: I did my best to sleep, and seem to have gotten about 4 hours. You might get redirected here to any other destination, in case you changed your original plan. Despite the fact that it’s apparently 3AM on my body clock, I don’t feel more than a little bit sleepy… yet. I’m sure it’ll catch up with me before I want to be asleep tonight.

My phone apparently isn’t working at all here: despite there being a recognized ‘T-mobile’ network, the phone just says “Access Denied” when the phone tries to join. Hopefully I can get that cleared up in Switzerland so that I can do some GSMLoc recording. There’s also no wireless here that I can find — there is an ad-hoc network here called “Free Public Wi-Fi”, but it doesn’t seem to respond to DHCP requests, so I don’t know what’s up with that. This is leaving me woefully underconnected — something I’m horribly unused to. In the US, I can *always* hop online, even if it’s slow, with putty on my phone, or connecting via Bluetooth to the laptop. Losing that ability is a strange feeling.

As far as Schipol airport goes, there are a couple things here that surprised me:
* How big it is. The D terminal has *87 gates* — and gate D6 (the one I’m at) has A-M subgates, which makes me think that there may be significantly more than that in total. How the heck that works, I don’t know, but it just seems big to me. Not to mention that there are 4 other terminals.
* Long taxiways! I think we were taxiing for about 15 minutes at full steam: we went from where we landed, in the middle of a bunch of fields, into what felt much more like a city. I’m used to taxiways being about hte same length as runways, so this was definitely a new one to me.
* Security-at-the-gate: Apparently, security here is specific to the gate, rather than to a wing of the airport. (Or maybe it’s both — I didn’t actually walk out.) Seems like a massive repitition of resources: I’ve never felt that Security for the terminal was a poor way of doing it: it seems like no matter what, one side is clean, and the other side is dirty: increasing the dirty area and decreasing the clean area does not seem to me like it would be important. (Granted, the dirty and clean sides don’t always stay that way, vis terror attacks that have happened or other incidents where security risks have gotten through — but that’s an issue of missed screening, not one where someone actually didn’t *get* screened.)
* Smoking in the building. I don’t think I’ll ever like this, given my significant difficulty breathing in smoking areas. I like Massachusetts and its “no smoking in bars” laws, thanks.

I’m also slightly amused by the airport warnings for late passengers: “You are delaying the flight. Immediate boarding please! We will proceed to offload your luggage.” Although Jess says that my argumentative, regimented, and strict behavior is ‘German’, I think that it has definitely bled over into Amsterdam as well. I wonder if that is true of more areas of Europe?

Altogether, other than the languages, flying here isn’t much different than flying to any other big airport. Then again, I haven’t gotten to my final destination yet 🙂 More reports as more things happen — for now, signing off, from Schipol-Amsterdam airport.

(This post was written at 2:39AM Eastern Time, 8:39AM localtime.)

At times, this weblog has been the source of information about Symbian-Python hacking, RDF hacking, and geohacking. This week is going to be a lot of the latter, as I’m travelling to Lausanne, Switzerland to attend and present at the Free Open Source Software for GeoInformatics conference, FOSS4G.

I’m typing this from gate D6M of Schipol-Amsterdam airport, as I wait for the connector flight from here to Geneva, Switzerland, where I will meet up with my coworker, Schuyler Erle, and we will take the train from there to Lausanne. This afternoon, we’ll be attending the open portion of the Open Source Geospatial Foundation face to face meeting. Tuesday and Wednesday are workshop days, during which we will probably recruit available OpenLayers hackers — Tuesday night there is a dinner gathering for OpenLayers users, developers, etc., although I don’t know that we’ve been real clear on where that is yet 🙂

My participation in this conference is being graciously sponsored by my employer, MetaCarta. MetaCarta has supported the development of OpenLayers, one of the up-and-coming tools used in the Open Source world as an alternative to the Google Maps API. OpenLayers is a stable and well-developed API, released under the BSD license, and originally developed by MetaCarta for business reasons, but was released to the public as it was obvious there was a need for such a tool in the open source world. The OpenLayers presentation will cover some basic history of the project, what it currently supports with demos and code examples, where the project is heading, and hopefully time for people to bring up questions about OpenLayers. This presentation will be at 8:30 on Friday morning, in Auditorium C.

I’ll also be talking a little bit about the MetaCarta Labs GeoParser API. The GeoParser API is a publicly available interface which demonstrates the functionality which MetaCarta tools can provide: namely, the ability to extract from documents the geographic locations referenced within them. This functionality is used in Gutenkarte, an application which allows you to browse books while referencing their locations to a map. The presentation for Gutenkarte and the GeoParser API is at 11:00 on Friday morning, in Auditorium C.

At the same time, we will be working with many of the people that we often work with, but in person, rather than online: this will be my first chance to make face to face contact with a number of important members of the geospatial community, and to repeat contact with some who I knew and met at Where, but didn’t understand the significance of. 🙂

If you’re interested in learning more about OpenLayers, Gutenkarte, for the GeoParser API, or would like something specific covered in a talk, please drop a line to labs@metacarta.com. MetaCarta Labs is the public R&D face to MetaCarta, and we would love to use tools which MetaCarta has to make more useful public APIs available. The only way to improve things is to receive feedback. What services would you like to build on top of the GeoParser API? What more could it do to make your life easier? MetaCarta is devoted to expanding the information available via public APIs, and helping open source users to build new cool tools on top of its APIs. If there is some limitation that is causing you to be unable to implement a service, we’d love to hear about it. MetaCarta is extremely supportive of Open Source projects, and the Labs branch of MetaCarta is the place to bring any ideas for what should be changed, or what new features could be added that would cause the services offered to be more useful.

Schuyler and I are attending this conference not only to present, but to converse: to find out what kind of tools MetaCarta can share that will benefit the community at large. We’d love to hear from anyone who wants to use these tools about how they can be made easier to use — and how you’re using them already.

You can find us in the FOSS4G IRC channel, at #foss4g2006, on irc.freenode.net, or email labs@metacarta.com. This is your chance to make yourself heard. We encourage you to take advantage of it.