Gav's Blog

Clearly following on from the antipodes, the UK Government is now holding a mashup competition as well. But it appears that they haven’t quite gone as far as we have in New Zealand – at least not in terms of trying to remove as many restrictions on the use of Government data.

Having a look at the data page, where the Government has published data, it is interesting to note that citizens are required to enter click-use licensing agreements, and commit to using API’s to access government data. You can see some of the frustration on the page listing the data downloads.

Why are they such a problem? Well, the data can only be used for non-commercial purposes – that is going to hinder development somewhat, with no potential return at the end of the work. The licensing agreements appear to allow the government agency to pull the plug at any point if they don’t like what you are doing – I wonder what would happen if a mashup showed an inconvenient truth for a government, would that be reason to get a government agency to pull the data plug? Another joy of accessing data through API’s is that they are limiting queries on their servers to a few thousand queries per day (in the case of the Ornance Survey).

Which is why Government data really should be properly freed. It needs to be released to obtain a life of its own under a suitable Creative Commons license. Let citizens download the data, mash it up on their systems, and deal with all the issues. Governments must not become the gatekeepers of this information through API’s and license agreements.

Having just been through the process of obtaining and publishing some Government data, I was interested to see Dennis McDonald covering an Australian blog post on eGovernment on ‘Making data freely available‘.

Craig makes some very good points in his conclusions, and I’ve got one that I would like to add based on recent experience with the State Highway trackpoint data I sourced in NZ and uploaded to OpenStreetMap.

In the followup to uploading the trackpoints to OSM, a group of NZ OSM mappers have been discussing how to classify roads in New Zealand – basically how the road hierarchy is classified. Turns out that this is a hard problem and there is not a single standardised approach taken to classifying roads in New Zealand. After asking for guidance, I became aware of a number of issues and capture these on the gis.org.nz wiki.

The simple fact was that there was not a nationally consistent approach to classifying the road hierarchy in New Zealand. I expect it is similar in Australia, and possibly worse with the additional layer of State government.

This highlights an additional key role that I think Government has to play in national data collection.

Government should set the standards for data collection to ensure that datasets are nationally consistent to enable simple aggregation of disparate datasets.

It doesn’t have to actually perform the aggregation, although that would be nice. It just has to ensure that standards are used to enabled aggregation. Using the road hierarchy classification example from above, this means that a road of a certain class in one part of the country, means exactly the same in another. This would probably occur by an engagement-based approach that determines a controlled vocabulary to define types of roads.

Without this responsibility, citizens are doomed to a million-and-one datasets that cannot be easily aggregated to produce coherent and consistent national datasets.

I’ve been meaning to blog about this sooner, but have been pretty busy with work. A chance email on a NZ GIS list that I belong to two weeks ago, inspired me to go out on a limb and see if I could get some Government data. I saw a post from someone within the Transit (soon to be merged into the New Zealand Transport Agency) refering to working with 2.2 million trackpoints from a roading survey. I started a private email discussion, and after a couple discussions, I soon had 2.2 million trackpoints from the 2008 High Speed Data Collection survey of New Zealand State Highway network.

My intention of obtaining this data was to be able to convert it to GPX files and upload it as a raw data survey layer to OpenStreetMap (OSM) so that it could be used as the basis for mapping New Zealand’s State Highway network in OSM.

I had some help from John McCombs from Integrated Mapping in Christchurch who very kindly reprojected all the points to WGS84. I then spent 4 evenings last week converting to GPX and uploading the files to OSM.

Was this data essential to mapping the highways in OSM? No. But it was a great experiment to see if a New Zealand Government Agency was willing to release data under acceptable terms and conditions – this dataset is licensed under the Creative Commons v3 Attritbution ShareAlike license, and effectively turn the raw data over for public consumption. Naturally, this doesn’t contain all of the detailed geometry that is collected during the survey, so not all of the data was made available, but we got the most important – latitude and longitude, and a lot of them!

For more information, see the following links.

• Transit High Speed Data Collection on gis.org.nz wiki
• Download the GPX on OpenStreetMap

One of the key points I was trying to make, was indicating that citizens are actually interested in accessing government data such as this, and that agencies should take a more proactive approach to releasing data for the world. After all, data is global these days – put it on the Internet and anyway can access it.

Over at the In Development blog at eGovt, they are asking for comments on the state of progress for eGovernment in New Zealand. I’ve copied my comment here.

Location, Location, Location!
A quick skim and keyword search indicates that the report has almost entirely neglected one of the most important aspects of information, and that it location – its spatial component. It is not surprising then that the Government is struggling with how to handle spatial information. Some of the most important aspects of community anywhere are distance and connectivity. However with Government doing a poor job in making spatial information available to the citizen, it is no wonder that individuals are struggling to find out about consents from two properties over.

Even more of a shame is the fact that Government has failed to deliver on a project to produce an authoritative National Address Register of addresses, roads and placenames. This dataset is one of the most fundamental to being able to place information in space under truly understand its context. eGovernment will never take off under the core location issues are worked out. Whilst the recent release of significant amounts of Statistics New Zealand information is a great step, the true benefits and insights contained in the released census data is being held back by the lack of an authoritative National Address Register.

Organisational Inertia and Champion Individuals
Government organisations themselves are often hesitant to change or try something new. This is not to say it doesn’t happen, and there have been some excellent examples recently of organisations stepping up and trying something new – such as using wikis for consultation and engagement. I am fairly certain that most of these projects will have been championed internally by a small group of individuals, and I’ll bet they had quite a struggle to see their project through. Here’s hoping that these recent projects are the catalyst for more risk taking, and acceptance of exciting IT projects within Government.

Not enough Champions
As a consultant that has worked  a lot within Government agencies, much of the information projects I’ve seen successfully delivered have been almost solely on the back of one or two key individuals. The agencies themselves have often been more of a hindrance to delivery of exciting new initiatives, and the Champions have had to fight an uphill battle against their own organisations, and other Government agencies when ‘engaging stakeholders’. Even worse, a number of these Champions actually run up so many brick walls in their organisation, and within Government, that they end up choosing to leave in frustration, or even worse are forced out by unenlightened superiors. In some circumstances they make their way to an organisation that is supportive, but that fate does not appear to wait all Champions. You need some way to support and encourage these key individuals that lie within Government agencies – there may only be one or two, but they are your key instrument of change from within.

Start Small, Release Early, Upgrade Often
Despite the fact that the National Address Register is such a necessary component of eGovernment. I don’t believe that the project as tendered should have gone ahead – not for something where tender prices ranged between $9 and $48 million. The project would have been so large and complex that it would likely have not been delivered on time, would have been over budget, and would have lacked the desired capabilities.

I think Government has to learn to try the small and simple things first and work up.

Surely, for the price of one or two Geospatial Professionals and supporting hardware, the New Zealand Government should be capable of aggregating all roading information from the 74 Road Control Authorities and Transit New Zealand, and publishing it as a single national roading dataset under permissive licensing. This should be able to be done now for a couple of hundred thousand dollars. But just produce a dataset and get it out there. Get feedback about how it is used, and look at improving the process, and accepting feedback. If you don’t, then the New Zealand Government is going to look pretty ridiculous when volunteers have created their own national roading dataset using OpenStreetMap because Government wasn’t capable.

Here are some simple actions that Government can undertake to encourage more exciting use of Government information.

Find data. Release it under permissive licensing. Release it in formats that make it readily accessible to manipulation in software (e.g. don’t release maps only as pdfs, make the underlying spatial data available). Announce it through a simple clearing house – nothing flash, it only needs to be  WordPress blog pointing to the relevant source. It won’t be until this happens that the more exciting concepts such as entirely unexpected but useful mashups occur.

And that is when things will get really interesting. At that point, we will have citizens building mashups and services about ‘Our Place In Space’ – and they won’t be constrained by the organisational inertia inherent in most Government agencies that are tied back by accountability and liability that makes them hesitant to take bold steps.

Remember eGovernment is not just about Government developing systems and solutions. It does include citizens, communities and organisations building systems to meet their own needs.

Note – this is a living document and will be continually updated with more observations as I spend more time using the unit.

Introduction
For a couple of years now, I have been borrowing a friends Garmin 60csx for going geocaching with. My last GPS was a Garmin GPS Map 76S, and before that the venerable 12XL. With my use of the 60csx steadily increasing, my friend was seeing less and less of their GPS, so I finally decided after some reading to take the dive and purchase one of the new Garmin Colorado’s to become my primary GPS.

This article is my living review of the Garmin Colorado 300 that I purchased, and I’ll be continually updating the page as I find more likes and dislikes. Any additions after the first publication will have the date of addition in brackets. I’ll mark positive and negative points with the obvious +/-.

Summary
I had heard quite a few comments and discussions about the Colorado’s, and had the brief opportunity to try one out up in Hamilton recently thanks to RadioNut and ButterflyLady. I had been tossing up between trying a Colorado and getting a Nuvi for the car, but the need to stop hoggin my friends 60csx made the decision for me.

To date I haven’t regretted it. The benefits so far appear to outweigh the hassles, and I’m sure that there will be many improvements made to the software in due course. Haven’t yet cached with it, but it looks promising – although I don’t think it will replace the PDA notes functionality I use for recording multi-waypoint information and calculating final caches.

System Software History
* Initial Writing – Software Version 2.40, GPS Software Version 2.60

Casing and Connectivity
The new exterior to the unit is remarkably different. Most noticeable is the big shiny screen, the roller, and the silver finish.

+ In the hand – the unit fits in my hand much better than the 60csx and feels much more solid and durable.

+ Big screen – more real estate taken up by the screen is a good thing!

+ External antenna port – this is in a far better position than the 60csx, and is far less likely to come out during use. The problem I had with the 60csx was the rounded casing around the external antenna port caused the plug to come loose with just the slightest antenna cable movements. Initial tests in the car show this port to be far better designed.

+ Latch system – the new latch also looks a big improvement over the 60csx. Having previously attempt to carabena the 60csx to something using the metal ring on the back, I quick discovered that with a big of movement the latch would soon open the battery compartment and you’d end up with the GPS on the ground, and just the back lid still attached to you. Not good. The new latch appears to secure the back much better.

+ Mounting system – the new mounting system on the back also seems a big improvement. I haven’t got the bike mount set up yet, and the car mount hasn’t arrived, but it appears to have a very nice rile system to slide the unit snugly into mounts.

- Silver finish – sure it looks high tech and all, but I’m wondering how long it is going to remain in that pristine condition once it has been on a few geocache hunts with me.

- USB Storage mode – why oh why did Garmin decide to have the Colorado automatically switch to USB storage mode as soon as a USB cable is connected? The 60csx was fantastic in that you could power the unit in the car via USB. Now you plug it in with the same cable and the unit switches to USB storage mode. Even worse, because you can’t eject it, you have to remove the batteries to reset the unit to its ordinary operating mode. For sure, make the storage mode easy to access when connecting it to a computer, but don’t have it default to it. Perhaps pop up a brief window that gives the option of converting to USB storage mode, but if no input is given in a few seconds then the GPS should stay in the same state.

Configuration and Profiles
The Colorado can be a bit of a monster to configure at first glance. It introduces profiles for different activities that allow you to quickly change the setup of the GPS.

+ Profiles – it is great being able to customise the the GPS for different activities. The unit came with profiles for Recreational, Geocaching, Automotive, Marine and Fitness activities. Each profile allows full customisation of all settings of the unit. Even allowing different co-ordinate systems, or measurement systems to be defined.

+ XML Profiles – I haven’t had much more than a casual poke-around on the USB file system of the GPS yet, but it appears that the profiles are stored in XML. This means in theory that you may be able to create a global XML template, and then adjust them in a software editor or even a web page. This may be particularly useful for sharing profiles with mates, or for suppliers to create profiles, and then upload new profiles just by copying the XML profile files into the appropriate folder.

- No global profile – one problem with the very configurable XML Profiles is that there doesn’t appear to be a global profile. This means that if you want to change a setting in all profiles you have to either load and edit each profile individually in the GPS, or change it in each XML file. What would be really nice is being able to have a Global profile (XML, just like the rest) that defines settings, and then each profile is able to either overwrite a setting, or inherit the property from the Global profile. Likewise, if the profiles are able to inherit properties from the Global profile, then the XML profiles would be smaller and easier to edit as they would only have to record the settings that are different from the Global profile, rather than record every setting.

Interface and Input Methods
This has been perhaps the biggest learning curve yet, as the Colorado has a significantly different user interface and input methods.

Having come from the 12XL, 76S, and more recently the 60csx, I have become very accustomed to the ‘page’ style of navigation.

+ After a few days usage, the roller is very very nice for changing the map scale, and sure beats the continue – in, in, in or out, out, out on the 60′s.

- Shortcuts Menu – I’m not yet convinced this is the easiest way to navigate around the pages, and for the time being I have changed this to the more familiar ‘page’ navigation. This can be changed in System Setup Page > Shortcuts > Options > [ Use Shortcut Menu | Use Page Loops ]. This may be where profiles come in. If using a profile with only a few pages, it may be better to use Page Loops. If using a profile with a large number of pages, the Shortcut Menu is probably a far better option. At least with profiles, you’ll have the best choice.

- Navigation strangeness – When navigating to a waypoint, and using Shortcuts, the Stop Navigation function is found under the Shortcuts list. However, when using Page Loop, it is dynamically added to the top of the Other list. Strangely enough, neither of the two obvious pages that may be used for routing – Map or Active Route – provide a quick option to either Stop or Detour the current navigation. Detour doesn’t even appear in the Other list in Page Loop mode. For some reason, when you are navigation, and are set in Page Loop mode, a second Mark Waypoint is added to the Other list, so you end up with Stop Navigation, followed by two Mark Waypoints. Clearly a bug with handling these dynamic menus. As it is currently structured, the Shortcuts mode appears to be far more reliable and easier to access Stop Navigation.

- Profiles – it should be far, far easier to change profiles. Right now you have to head to either Setup or Other, select Change Profile and select the profile. Too hard for constant changing, particularly if you’re going from Automotive to Geocaching as you drive to a cache, hunt a cache, drive to the next cache etc. It should be possible to load profiles into the Shortcuts menu.

- White message pages – these have frankly become a little irritating. Having the whole screen revert to white to display an error message is annoying. Particularly when you are driving and navigating and you are not necessarily in the possible to dismiss the error. More subtle error messages should be implemented that allow you to continue to at least see the display of the GPS.

Function – Tracks
The tracklog is a mixed bag so far with the Colorado, and I haven’t yet quite got my head right around it. The setup is relatively simple and allows you to select Auto/Time/Distance as always.

One thing I loved about the 60csx – particularly for travel – was the way you could set it to record track points quite frequently, and have it save them to the card. Each day it would create a new tracklog file with the date as the filename – e.g. 20080508.gpx. This was great.

The Colorado setup is different. In particular is the way it handles Current vs Archive tracks. This of course may just be nomenclature in that the Current track is the same as the 10k active track on a 60csx, and the Archive tracks are the equivalent of those save to a card. However, instead of naming and filing the tracklog GPX files by date, the Colorado appears to be using a serial number e.g. 1.gpx, 2.gpx etc. It appears to create a new file at a certain size, rather a separate file for each day.

Function – Geocaching
The geocaching functionality on the Colorado certainly deserves a lot of discussion. Some aspects of it work very well, others leave a lot to be desired.

+ Geocaching GPX – the ability to store geocaching GPX in the unit and display cache information is quite useful. It provides nice quick access to difficulty, terrain, cache size, trackables, full description, hints and logs. This certainly brings the Colorado a long way towards true paperless geocaching.

+ Field Notes – these appear to work very well. After heading out for some geocaching where I found four, and DNF’ed another, I tested the upload of the field notes file. It worked flawlessly and provided me with an easy means of creating logs for each cache, and these were removed from the list as the caches were logged. The are loaded in the order they were logged in the field – making the logging of caches easier than ever before. I am certainly going to make good use of this function. Of course, given that field notes are nothing more than a text file, it could also be crafted on a PDA in the field and uploaded. But the Colorado certainly makes it painless.

- Geocaching GPX for multis and puzzles – the geocaching GPX is fantastic for simple geocaches such as traditionals and letterboxes. However, it does fall down when multicaches are concerned. Multis still required something to record details and solve to find the final cache. Coming from a PDA, I still found myself reaching for the PDA to complete the multi-caches today.

- Geocache waypoints – Unfortunately the geocaching functionality does not recognise multicache waypoints, or carparks and the like. Instead of finding them under the Geocache page, you have to return to the Where To? page, and go to waypoints and then try and find the waypoint. The Colorado really should be smart enough to match the 3rd-6th/7th characters of a waypoint name so that it can match it to a geocache, and then display these child waypoints under the Options menu with a nice quick Go To Child Waypoints page or similar.

- Routing – it is quite cumbersome to  select road routing to drive to a cache, and then revert to geocaching mode to find the cache. All navigation in the built in geocaching mode is point-to-point. The Geocaching profile really needs an additional option when selecting Go To Location – it should ask whether you would like to route or go direct. I found it quite frustrating today to switch back and forwards between modes today so that one could route from cache to cache in the car.

- Cache identification – it would be nice on the Geocache page if the type of each geocache was identified with a little icon. It would help when deciding what cache to do next to see whether it is a traditional, multi or puzzle.

The Geocaching functionality has a lot of potential, but it is still in a relatively immature state. The capabilities are huge step up from the 60csx in terms of complexity, however the ease of use and thought about the interface doesn’t look that refined.

Function – Automotive
Wow – used the Automotive profile today and was very impressed by the map. A nice oblique view looking down from behind to show the streets ahead. Upcoming streets a nicely labelled. I’m really looking forward to getting the car mount now and using it. It looks very nice.

Function – Satellite
It is a simple page, but one I refer to quite often to check the signal. Whilst the display of information is very nice, I feel that it has a few improvements that could be made.

- Signal Strength Bars – the dark blue signal strength bars are very difficult to see against the default dark green background. Please provide a higher colour contrast setting so that you can actual see the signal strength bars at a quick glance.

- Unknown Number – there is a little number down the bottom right of the constellation map, but it is not identified. I assume it is the elevation. Please mark it as such.

Function – Map
What can I say other than the new map looks absolutely fantastic. Most of these comments are based on the NZ Open GPS Maps and the Global DEM.

+ Global Digital Elevation Model (DEM) – the inclusion of a global DEM means that even New Zealand to some degree gets the pretty shaded relief map imagery when zoomed out. Note that this is not the 3D view, and that requires additional information to be added to the maps to support that.

+ Resolution and Screen Size – the combination of increased screen size and increased resolution have created a gorgeous and practical display. Practical as you can display more detail on the map by using smaller cartography.

+ Tracklog Mapping – the mapping of tracklogs is far less intrusive than it used to be. A nice thin black line that doesn’t obscure the actual roads is a big improvement. It even allows you to see more detail in you tracklog, such as the line around roundabouts. Very nice.

+ Options – the map options are very simple and easy to use. It is simple to select maps and enable/disable them with just a few simple clicks. Likewise adding and removing data fields are quick and easy. Modifying data fields couldn’t be much easier either.

- Only Two Data Fields – with the increase in map size, it might be nice to have say four data fields on occasion, but the options didn’t appear to allow anything other than the default two.

Function – Compass
I haven’t really used the compass yet, so more comments will come.

- Easy On/Off – it seem really strange that to turn the compass on and off, you can’t do this from Compass Page > Options, rather Setup > Heading > Compass > [ Off | Auto ]. You should be able to quick toggle this from the Compass page options.

Function – Project Waypoint
This is a really bad example of user interface design. I can’t believe that to enter the bearing, you have to enter each digit of the degree. How slow, painful and indicative of someone that hasn’t really thought about the design. They should use a different graphic to indicate a compass, and have fast turns translate to +/- 10 or 15 degree increments, and have slow turns refine the bearing by individual degrees. Far quicker entry than the tortuous system they have now.

Recommended Improvements
These are listed in no particular order, and are a brief summary of what is contained in the review above. I’m sure some of these we will see in time with software upgrades – assuming that Garmin just wanted to get the Colorado’s to market, and then continue adding/refining features to an existing userbase.

* Create a global profile, and allow custom profiles to inherit or override properties from the global profile. This will make profile management quite a bit easier.
* High contrast signal strength bars on the Satellite Page.
* Identify the mystery number of the Satellite Page.
* Allow the display of more data fields on the map display e.g. default is two, allow say four, or even six.
* Provide a quick Compass Off/Auto option on the Compass Page.
* Provide a different user interface for selecting degrees on the Project Waypoint page. There is absolutely no reason to enter each digit individually for a bearing.
* It would be nice on either the Satellite or Map pages to have an option to quickly turn the Tracklog on/off.
* Make it easier to access the Stop Navigation/Detour modes when in Page Loop mode. In fact, why not just put these under the Options menu for both the Map and Active Route pages – as these are the most commonly used pages for routing.
* Profiles should be able to be added to the shortcuts menu to speed switching e.g. I’d like to be able to quickly select Automotive and Geocaching profiles with as few button presses as usual.
* Geocaches page needs to identify different cache type at a glance.
* The Geocaching function needs a much easier means of integrating routing for driving between caches, and then switching to cache hunting.
* Stop the unit switching to USB mode when a USB cable is plugged in.  Why can’t it work just like the 60csx? USB should power it by default, and perhaps the USB mode could be selected from the Shortcuts instead
* Get rid of the full white page error messages.

Just wanted to add some brief comments following discussion as GOVIS 2008, in the ‘Our Place in Space’ stream. This was held in Wellington last Friday. At the end of the day I raised the following challenging comments to the audience, and they seemed to be well received. I will be reviewing and adding to these in the near future.

1. Don’t wait for application ideas or business cases. Free as much data as possible and the applications will follow. The killer application for your data has not been invented yet.

2. Consider community and volunteers. We are enthusiastic, have a huge capacity to learn, we have a hunger for data, and if it suits us we will dedicate massive amounts of time to it. Our expectations are being fed by online mapping and we want more.

3. Spatial citizens don’t care about arbitrary political boundaries and are happy acting with national and even international datasets. We don’t want to work with 80+ Local Government agencies. We want it to be trivial to find spatial information. Let us integrate the data we want using our own systems. We want to package the data our way. We want to take it out into the real world where there is no web access. We want to fix your data, but only if it is fixed once and fixed fast. LINZ – are you listening? We can do it in 24 hours and we’re not getting paid to do it. We will only get faster, more competent, and hungrier for information about our place in space…

As part of the GOVIS Geospatial miniConference, a maps mashup is being held the day before with data sets being provided by LINZ and others. More information will be made available on the barcamp page (for more on what a barcamp is – click here).

A challenge to innovate! A challenge to find open data! Create and present your mash-up with a few data sets provided for the BarCamp! Cool Prizes! Sponsored by Statistics New Zealand, The New Zealand Geospatial Office and the Spatial Sciences Institute. Entry is open to everybody who is enthusiastic about using New Zealand’s core geospatial data in presenting current issues and analysis challenges! MashUp 2008 is an event which brings together New Zealand’s leading technical experts, as well as budding enthusiasts, in combining information sources with mapping boundaries and data in innovative ways. Rules of the competition will be downloadable here as soon as possible.