Friday, 18 August 2017

Making GovHack (and Open Government) more impactful

I've finally attended the GovHack weekend. As a dad, weekends used to be for taking kids to birthday parties and soccer games. But my boys have grown up, giving me the chance to see how GovHack compares to the Open Source communities I've been involved with for decades. I wanted to see what each can learn from the other and signed up as a coach.
GovHack is an annual event where volunteers band together for 48 hours to write applications with Open Government data. Participants compete for prizes for the most innovative and useful applications. It has grown every year since it started in 2009, attracting thousands of volunteers, running in 36 locations across Australia and New Zealand, and attracted numerous sponsors and an excessive list of open government datasets. Credit must go to the organisers for creating such a sustainable winning formula. But lets ask some tough questions and hopefully help GovHack become more impactful in future?

What is the point of GovHack?

What is the point of GovHack? It wasn't obvious from looking at the main website, but I found an answer buried in the GovHack 2016 Year in Review:
In his opening address, Craig Laundy, the Assistant Minister for Industry, Innovation and Science highlighted that open data was one of the keys to the Australian Government’s National Innovation and Science Agenda. He read a letter from Prime Minister Malcolm Turnbull which paid the following tribute to Govhack:
“Data without ingenuity is like a lamp without power – only when the two are connected do opportunities to innovate become clear. This is why GovHack is so important.”
Recommendation 1: We should be clear about the purpose and value of GovHack. We should prominently promote messages like "GovHack aims to contribute to the government's Innovation Agenda by encouraging and facilitating ingenuity with government's open data."

Is GovHack enabling Innovation?

So how successful has GovHack been at enabling innovation? It's hard to say really. The 2016 Year in Review provides plenty of details about numbers of participants, datasets used, awards, VIP presenters, red carpet events, but there is barely a mention of how successful GovHack has been at enabling innovation. The best I could find was a passing mention of an "IP Nova App" which started in GovHack 2015. I’ve since been told about a couple of others. But the point is that we are measuring how busy everyone is, and how much buzz is being created, but completely failing to report the impact on innovation.

Recommendation 2: Let's measure and report on the realised innovation resulting from GovHack. Let's then assess results and work out ways to improve GovHack's impact on innovation.

Maturing ideas is hard work

Why is it hard to find reports of GovHack ideas progressing into sustained initiatives? I can't say for sure, but suspect very few GovHack ideas actually grow into something. The simple truth is that good software takes substantial effort to design, write, test, deploy and maintain. While a 48 hour GovHack is useful for brainstorming ideas, it stills requires significant follow up if it is to mature into something useful. And here we notice the difference between Open Source Code Sprints and GovHack. On completion of Code Sprints, there are established and experienced communities committed to adopting and advancing worthy ideas. Who in the GovHack community is offering to help take good ideas through to maturity? I don't see such support mentioned in GovHack web pages.

Recommendation 3: GovHack sponsors' should aim to realise true value by helping to mature innovative ideas into reality. 

The majority of people I saw in the Sydney GovHack appeared to be University students or recent graduates. For these young people, GovHack provides a great practical learning experience, some mentoring, and an opportunity to network. However I couldn't help feeling there was an level of exploitation of these young volunteers. Government agencies are gaining significant value from volunteers testing their datasets, something that would cost orders of magnitude more if implemented internally. Morally, I feel these agencies should give more than a free meal and a chance to share in a prize. A good symbiotic relationship would hopefully consider providing more value for our young community.

Recommendation 4: Sponsors should consider formally setting up cadetships or project development opportunities as awards.

How good is the data?

Integrating data into innovative web or mobile applications typically should follow standard design patterns, with data published through a web service, then processed, integrated, and presented in innovative ways. Ideally government agencies should make data really easy to use, setting up data web services and providing clear documentation and examples. Instead teams were spending much of their GovHack time setting up the infrastructure to publish this data rather than spending their time being innovative.
It is worth being reminded of one of The Australian Digital Transformation Agency Design Principles:
Principle 4. Do the hard work to make it simple.
Making something look simple is easy. Making something simple to use is much harder - especially when the underlying systems are complex - but that’s what we should be doing. Don’t take “It’s always been that way” for an answer. It’s usually more and harder work to make things simple, but it’s the right thing to do.

Recommendation 5: Government should define a best practices guide for publishing data services, and then follow this guide.

How does government know if they are doing a good job? Ruthless survival of the fittest principles apply to Open Source and market economies. People don't buy substandard products. Only the best Open Source projects attract communities. Again, refer to the DTA design principles:


Principle 5. Iterate. Then iterate again.The best way to build good services is to start small and iterate wildly. Release minimum viable products early and test them with actual users; move from Alpha to Beta to Live adding features, deleting things that don’t work and making refinements based on feedback. Iteration reduces risk: it makes big failures unlikely and turns small failures into lessons. If a prototype isn’t working, don’t be afraid to scrap it and start again.

Recommendation 6: Agencies should measure the usability and usefulness of their datasets, assess and adjust accordingly. GovHack provides an opportunity to measure these metrics.

How good are we are implementing Open Government?

And so I come to my most pointed point, which was recorded as a video for my GovHack contribution:


Australia has embraced great policies around Open Government. These describe how openness and collaboration enable innovation. However, the practical implementation of these open principles have proven elusively difficult, with reported success stories coming from a few charismatic champions rather than being systemic across all government.
Why is that? Well, it’s complicated. There is a wealth of established wisdom, spread across the domains of Open Source Software, Open Standards, Open Data, Open Government, and more. However, we still lack clear and definitive guides which draws all this wisdom together into practical playbooks which can be easily applied by government agencies. Instead, current government practices and guidelines regularly hinder collaboration. Let’s fix that.

Recommendation 7: Let’s build an Open Government Playbook.

Let’s document the subtle magic which makes open and collaborative communities work.
This Playbook should cover technology, processes, governance, leadership, business paradigms, and ethics. It should be written in simple language, designed to support decision makers, architects, implementers and citizens to understand open principles.

Could GovHack be more impactful?

Acknowledging that GovHack runs impressively efficiently and has attracted a huge ground swell of interest and momentum, could we make it more impactful? I think we can. We should remind ourselves of the Open Government and GovHack goal of promoting innovation. We should measure innovation enabled and adjust accordingly. Adjustments will likely include aligning more closely with Open Source development practices.

Thursday, 10 August 2017

OSGeo-Live 11.0 Reboot



Version 11.0 of the OSGeo-Live GIS software collection has been released, ready for FOSS4G which is the International Conference for Free and Open Source Software for Geospatial - 2017 in Boston, USA.

Download

Download the OSGeo-Live 11.0 image at http://live.osgeo.org/en/download.html

Release Highlights

This release has been a major reboot, with a refocus on leading applications and emphasis on quality over quantity. Less mature parts of the projects have been dropped with a targeted focus placed on upgrading and improving documentation.
Dropped
  • Windows-only applications/installers
  • Overviews of OGC Standards
  • Some applications that did not meet our review criteria
  • We now only support a 64 bit distribution (32 bit is built but not officially supported)
Added
  • Support for isohybrid ISO images with UEFI

Known Issues and Errata

Post release issues are listed here: https://wiki.osgeo.org/wiki/Live_GIS_Disc/Errata/11.0

About OSGeo-Live

OSGeo-Live is a Lubuntu based distribution of Geospatial Open Source Software, available via a Live DVD, Virtual Machine and USB. OSGeo-Live is pre-installed with robust open source geospatial software, which can be trialled without installing anything.
It includes:
  • Close to 50 quality geospatial Open Source applications installed and pre-configured
  • Free world maps and sample datasets
  • Project Overview and step-by-step Quickstart for each application
  • Lightning presentation of all applications, along with speaker's script
  • Translations to multiple languages
Homepage: http://live.osgeo.org
Download details: http://live.osgeo.org/en/download.html
Credits
Over 180 people have directly helped with OSGeo-Live packaging, documenting and translating, and thousands have been involved in building the packaged software.
Developers, packagers, documenters and translators include:

Activity Workshop, Alan Boudreault, Alex Mandel, Alexandre Dube, Amy Gao, Andrea Antonello, Angelos Tzotsos, Anton Patrushev, Antonio Santiago, Argyros Argyridis, Ariel Núñez, Astrid Emde, Balasubramaniam Natarajan, Barry Rowlingson, Ben Caradoc-Davies, Benjamin Pross, Brian Hamlin, Bruno Binet, Bu Kun, Cameron Shorter, Dane Springmeyer, Daniel Kastl, Danilo Bretschneider, Dimitar Misev, Edgar Soldin, Eike Hinderk Jürrens, Eric Lemoine, Erika Pillu, Etienne Dube, Fabian Schindler, Fran Boon, Frank Gasdorf, Frank Warmerdam, François Prunayre, Friedjoff Trautwein, Gabriele Prestifilippo, Gavin Treadgold, Gerald Fenoy, Guillaume Pasero, Guy Griffiths, Hamish Bowman, Haruyuki Seki, Henry Addo, Hernan Olivera, Howard Butler, Ian Edwards, Ian Turton, Jackie Ng, Jan Drewnak, Jane Lewis, Javier Rodrigo, Jim Klassen, Jinsongdi Yu, Alan Beccati, Jody Garnett, Johan Van de Wauw, John Bryant, Jorge Sanz, José Vicente Higón, Judit Mays, Klokan Petr Pridal, Kristof Lange, Lance McKee, Larry Shaffer, Luca Delucchi, Mage Whopper, Marc-André Barbeau, Manuel Grizonnet, Margherita Di Leo, Mario Carrera, Mark Leslie, Markus Neteler, Massimo Di Stefano, Micha Silver, Michael Owonibi, Michaël Michaud, Mike Adair, Milan P. Antonovic, Nathaniel V. Kelso, Ned Horning, Nicolas Roelandt, Oliver Tonnhofer, Patric Hafner, Paul Meems, Pirmin Kalberer, Regina Obe, Ricardo Pinho, Roald de Wit, Roberto Antolin, Robin Lovelace, Ruth Schoenbuchner, Scott Penrose, Sergio Baños, Sergey Popov, Simon Cropper, Simon Pigot, Stefan A. Tzeggai, Stefan Hansen, Stefan Steiniger, Stephan Meissl, Steve Lime, Takayuki Nuimura, Thierry Badard, Thomas Gratier, Tom Kralidis, Trevor Wekel, Matthias Streulens, Victor Poughon, Zoltan Siki, Òscar Fonts, Raf Roset, Anna Muñoz, Cristhian Pin, Marc Torres, Assumpció Termens, Estela Llorente, Roger Veciana, Dominik Helle, Lars Lingner, Otto Dassau, Thomas Baschetti, Christos Iossifidis, Aikaterini Kapsampeli, Maria Vakalopoulou, Agustín Dí­ez, David Mateos, Javier Sánchez, Jesús Gómez, Jorge Arévalo, José Antonio Canalejo, Mauricio Miranda, Mauricio Pazos, Pedro-Juan Ferrer, Roberto Antolí­n, Samuel Mesa, Valenty González, Lucía Sanjaime, Andrea Yanza, Diego González, Nacho Varela, Mario Andino, Virginia Vergara, Christophe Tufféry, Etienne Delay, Hungary, M Iqnaul Haq Siregar, Andry Rustanto, Alessandro Furieri, Antonio Falciano, Diego Migliavacca, Elena Mezzini, Giuseppe Calamita, Marco Puppin, Marco Curreli, Matteo De Stefano, Pasquale Di Donato, Roberta Fagandini, Nobusuke Iwasaki, Toshikazu Seto, Yoichi Kayama, Hirofumi Hayashi, Ko Nagase, Hyeyeong Choe, Milena Nowotarska, Damian Wojsław, Alexander Bruy, Alexander Muriy, Alexey Ardyakov, Andrey Syrokomskiy, Anton Novichikhin, Daria Svidzinska, Denis Rykov, Dmitry Baryshnikov, Evgeny Nikulin, Ilya Filippov, Grigory Rozhentsov, Maxim Dubinin, Nadiia Gorash, Pavel, Sergey Grachev, Vera, Alexander Kleshnin, kuzkok, Xianfeng Song, Jing Wang, Zhengfan Lin, Jakob Miksch

Sponsoring organisations


Tuesday, 30 May 2017

What's in the OSGeo-Live 11.0 reboot?

In line with the OSGeo-Live reboot we have identified the following actions:

We will only support a 64 bit distribution, (32 bit will be built but not tested or officially supported).

We will no longer provide:
  • Windows installers,
  • Write ups for OGC Standards,
  • Write ups for libraries without a user interface (on a case-by-case basis)

Stray projects:

We haven't got a point of contact for these projects. If you know something about this project, and can help review docs, then please consider adopting the project. Without a volunteer, we will need to consider retiring them from OSGeo-Live. Work commitment is low, kudos and karma very high :)
  • openlayers
  • osgearth
  • gmt
  • iris
  • jupyter
  • JOSM ( OpenStreetMap editor )
  • proj4
  • mb-system
  • zygrib

Projects recommended to be retired from OSGeo-Live:

  • geokettle
  • geomajas
  • javaworldwind
  • kosmo
  • mapnik
  • pywps (quickstart broken)
  • sahana
  • tinyows
  • ushahidi
  • viking

Remaining projects:

  • Most projects have at least minor tweaks required for the docs, some have outright errors.
  • We are asking each project to, check their project overview, run the application on an OSGeo-Live VM (10.5 or a recent nightly build), run every step of the quickstart, update if required, and tell us when done.
  • Once verified, we will add the project back into the OSGeo-Live 11.0 builds.
  • We will call on all projects to test their projects and quickstarts again in an OSGeo-Live 11.0 beta.
For specific project comments, see the OSGeo-Live status spreadsheet, column T "11.0 Comment", and column U "10.5 Doc Review".

Key Milestones

  • 5-Jun-2017 OSGeo-Live Feature Freeze (final application versions installed)
  • 19-Jun-2017 OSGeo-Live delivered to UAT (final application versions installed - Beta stage)
  • 24-Jul-2017 OSGeo-Live Final ISO
  • 14-Aug-2017 FOSS4G 2017 Boston
... full schedule

About OSGeo-Live

OSGeo-Live is a Lubuntu based distribution of Geospatial Open Source Software, available via a Live DVD, Virtual Machine and USB. You can use OSGeo-Live to try a wide variety of open source geospatial software without installing anything.

Tuesday, 9 May 2017

What applications and versions should be included on OSGeo-Live 11.0?

We are asking all OSGeo-Live projects to let us know:
  1. What versions of projects should be on OSGeo-Live?
  2. Are there any new projects which should be on OSGeo-Live?
  3. Are there any projects which should be dropped from OSGeo-Live?
As per last announcement [1], with our next OSGeo-Live release we intend to focus on quality, and as such will be asking each project:
  1. To help install the latest stable version of the software.
  2. To verify Project Overview and Quickstarts are both up to date and in line with our documentation standards.
You can see results of our first pass review in our status spreadsheet [2], column T "11.0 comment", and column U "10.0 Doc Review".

[1] https://wiki.osgeo.org/wiki/Live_GIS_Disc_Press_Release_75

[2] https://docs.google.com/spreadsheets/d/1Q5BaEgQtgw4O1bXyeWMlM8XtAOhUgcjZ7Y2O0FZc2H0/edit?hl=en_GB#gid=2014800150

Key Milestones

15-May-2017 Decide versions of applications to be installed on OSGeo-Live
22-May-2017 Draft installers for new applications complete
5-Jun-2017 OSGeo-Live Feature Freeze (final application versions installed)
19-Jun-2017 OSGeo-Live delivered to UAT (final application versions installed - Beta stage)
24-Jul-2017 OSGeo-Live Final ISO14-Aug-2017 FOSS4G 2017 Boston

... full schedule

About OSGeo-Live

OSGeo-Live (http://live.osgeo.org) is a Lubuntu based distribution of Geospatial Open Source Software, available via a Live DVD, Virtual Machine and USB. You can use OSGeo-Live to try a wide variety of open source geospatial software without installing anything.

Sunday, 7 May 2017

Building an Open Government Multistakeholder Forum

Background

The Australian government's Department of Prime Minister and Cabinet is asking for community input on the development of an Open Government Multistakeholder Forum: a group that will allow government and the community to ensure the commitments of Australia’s Open Government National Action Plan are realised. The plan's commitments involve strengthening and improving:
  • Transparency and accountability in business,
  • The availability of open data and the digital transformation of government services,
  • Access to government information,
  • Integrity in the public sector, and
  • Public participation and engagement.

Framing the question

Let’s be more specific by defining the question this forum should aim to solve:
How can decision makers make good decisions about complex topics which address community needs and reflect community values, incorporating community engagement and earning community trust?
Unfortunately, the utopian vision of open government has inconvenient practical limitations requiring a healthy dose of pragmatism in order to be effective. The following challenges exist:
  • How do you attract the attention and contributions from people, acknowledging that we are in an era of information overload and competing priorities?
  • How do you facilitate conversations about complex and intricate issues, acknowledging that it requires substantial time to fully understand and then debate concepts?
  • How can you assess comprehensive community opinion on complex subjects, acknowledging that most people don’t have time to engage in the study or debate.
  • How do you avoid inappropriate bias from vested interests, acknowledging that vocal minorities (companies or individuals) may promote self serving principles contrary to the interests of the majority?
  • How do you gain trust from communities in presented information?
  • How do you manage communication overhead from a large community, finding and promoting the best ideas? (Increasing the signal-to-noise ratio).
  • How can people with a good ideas know they will be heard, and hence be encouraged to participate, noting the best ideas for complex subjects will be time consuming to develop and express? (Increase voice-to-signal ratio).
  • How do you efficiently avoid conversations being derailed by inappropriate or disrespectful behaviour? (Don't feed the trolls)

Learn from Open Source

In answering these questions we can learn from the communities behind established Open Source Software projects. Most of the principles behind Open Government are inherited from Open Source best practice, which in turn are based on morals found in gift economies such as family gatherings and local community groups.
Established Open Source communities typically have embraced Merit-ocracy, Do-ocracy, Reciprocity, Respect, Modularity, Pragmatism, Sustainable Practices, and have attracted Strategic Funding. Each of these points should be considered in detail as to how they should be applied in the Open Government context.
Open Source projects typically are started by a small volunteer base. All the team know each other intimately. Leaders organically evolve, based on meritocracy and do-ocracy. If the project successfully grows, it builds a technical base which becomes too big to maintain on volunteer labour alone. The project either hits a glass ceiling, or attracts external funding to resource core activities (such as project coordination). Australian Open Government has grown to a size comparable to large established Open Source projects and should be managed and resourced accordingly.

Answering Specific Questions

Government have suggested a basis for founding an Open Government Multistakeholder Forum and asked for community input. Highlights include:
  • Purpose: [The scope of work described for forum members, covering review, building a reasoned opinion based upon analysis, debating and then explaining opinions as recommendations which would require a non-trivial amount of time for diligent members to complete properly.]
  • Ways of Working: The Forum [will] meet at least every two months, in a location rotated between capital cities.
  • Ways of Working: Forum members [will] not be remunerated, but that its community members be reimbursed by government for reasonable travel costs.
  • Structure: The Forum [will] comprise not more than 16 members ... with equal representation from government ... and the community.
  • Appointment Criteria:  The Forum [should] broadly reflect the [social] diversity of the Australian community. In particular ... women and men should hold at least 40 per cent of positions on the Forum. ...
Question 1: Are there any other functions the Forum should usefully perform?
Question 2: If you do not agree with the single forum model, how do you think another model should work?
Question 3: Are there other ways of working you think the Forum should usefully adopt?
Question 4: Are there any other criteria or guidelines that should inform the appointment of community members to the Forum?

Answers 1-4:
There is a significant mismatch between the apparent workload, the accountability that should exist, and the expectation that leadership should come from unfunded volunteers. This should challenge our underlying human morals and ethics.
  • For forum members to provide valuable leadership, they should allocate substantial time analysing, debating and deciding on issues. This would equate to multiple days per week, which is more than most volunteers have available to them. Note that you want to attract committed people already playing leadership roles in open project(s).
  • As well as being morally questionable, I'd expect that having the Australian government expect forum members to work for free is counter to Australian labour laws.
  • I would expect that nominated government employees would be paid to participate as part of their daily work. This will create a mismatch in effort and influence that can be applied by government employees verses community members.
  • If community members are instead paid to participate by an external employer, one should question likely conflicts of interest.
  • If continuing with unpaid forum membership, the scope of tasks expected of forum members should be reduced to be in line with typical capacity of volunteers. 
  • The "Purpose" for the Multistakeholder Forum should be accompanied by realistic time and task commitment expectations of forum members, such that candidates and the community have a clear understanding of what should be done.
  • As well as aiming for social diversity, the forum would benefit from having a multi-disciplinarian team. 
Question 5: How should a selection panel to recommend Forum appointments be composed?
Question 6: Should nominations to the Forum be published?
Question 7: Having regard to the desirability to appoint Forum members without unnecessary delay, is there a better way to administer the nominations and appointment process?

  • If forum membership is treated as a paid position, then candidate selection should be treated with the same privacy applied to standard job hiring.
  • If membership is unpaid, then a public process could be considered. For instance, each candidate could provide a brief statement about themselves. The community can then vote on potential candidates to help short list candidates. The final selection should go to the selection committee who considers community votes when doing face-to-face interviews.
Question 8: Should appointments to the Forum be staggered?
Question 9: How should any mid-term vacancies be filled?

  • More relevant is "How should the Forum be refreshed?" In order to ensure history from one panel to the next, only a maximum of half the panel should be updated in any election cycle.
Question 10: How can the Forum best hear and respond to the range of community perspectives on open government?

This question requires construction of a well thought out plan. Here are a few memory jolters (to be considered with comments above):
  • Define a clear vision and set of principles which casual members of the community can quickly assess and agree or disagree with.
  • Trace all actions back to these principles.
  • Acknowledge the diversity in your community's level of interest and time commitments and adjust accordingly. 
  • Make use of tools which efficiently capture crowd wisdom. Eg: Allow comments on posts, and allow communities to vote up/down comments based on merit.
  • Respect the time and value of your community. Treat them as partners rather than a survey point. Adopt tools and processes which enables the community to maximise each individual's value of each contribution.

Post Note

This response is based on my background working within Open Source communities, Open Standards communities, activist communities and local communities. Do these ideas make sense to people with different experiences?

Tuesday, 18 April 2017

Reducing scope of OSGeo-Live for next 11.0 release


For our next OSGeo-Live release, 11.0, we propose to reduce the number of packages included, and only support a 64 bit distribution, (32 bit will be built but not tested or officially supported).
Factors leading to this suggestion include:
  1. Some projects have dwindling communities and momentum.
  2. Increased OSGeo-Live scope has increased our core maintenance and testing.
  3. Reduced engagement from projects (partly due to less core time spent reaching out to projects)
  4. Missing our first release milestone in 9 years.
From our options of reduce quality, become more efficient, increase volunteer engagement, find a sponsor to support core activities, and reduce scope, reducing scope is our most viable and acceptable option. Other ideas are welcomed.

Questions we will ask in assessing which projects to keep include:
  1. Is there an ACTIVE OSGeo-Live liaison person/people for the project? See "Contact" column in our Package List.
  2. Has the Project Overview and Quickstart been reviewed and are they current and complete?
  3. Do OpenHub metrics reflect an active and healthy community: 
  4. Is the project being updated on OSGeo-Live with each release?
Key Milestones
  • 5-Jun-2017 OSGeo-Live Feature Freeze (final application versions installed)
  • 19-Jun-2017 OSGeo-Live delivered to UAT (final application versions installed - Beta stage)
  • 24-Jul-2017 OSGeo-Live Final ISO
  • 14-Aug-2017 FOSS4G 2017 Boston