FOSS4G:UK 2022 Local at the British Geological Survey

Increasing open source adoption at the British Geological Survey

⚠️ Warning: bias ahead ⚠️

About open source software

Image from [choosealicense.com](https://choosealicense.com). Free and open source licences emphasise freedom. Users are free to use the code for anything that they want. The most important thing is to acknowledge the original authors. This includes Mastodon (open source Twitter equivalent) being used for Donald Trump's truth.social. The MIT licence is most permissive and only requires users to include the licence notice and acknowledge source. GPL3 is recursive and requires code linking to these files also to be released with a GPL3 licence. This is the preferred option of the Free Software Foundation because it ensures that everyone benefits. The AGPL licence is like the GPL3 but applies to software running on a server. It was created to prevent cloud vendors selling products based on open source without contributing anything back.

Advantages of open source software

Open source software is written to get things done; the purpose of proprietary software is to make money. These annoying click-bait links on the Windows lock screen will open in the Edge browser, displaying results from Bing search. This gives Microsoft a chance to push you towards their products. No-one asked for this and it doesn't help get things done, but it does make money. If someone tried this in an open source project, someone else would create a fork without the ads and people would use that instead. FOSS emphasises solving problems over making money. Code is written to solve the problem, so there is no problem giving it away. If code is written to make money, then some kind of licence key is required to see who has paid. This leads to all kinds of extra ways that the software can stop working: + Have no licence + Licence has expired + Too many others are using the licence + Don't have licence dongle with me + Licence server unreachable + Licence server down None of these issues exist with open source software. Image source: https://answers.microsoft.com/en-us/windows/forum/all/lock-screen-has-no-login-dialogue/b59ab023-f4b3-406f-8066-f792af55c25c

As a developer, you can benefit from many solved problems - Standing on the Shoulders of Giants. We added the CQL feature to pygeoapi. We did this because is was functionality that we required. We already used the normal version of pygeoapi and were benfiting from the work of others. By adding this feature, others benefit from our work. The webpage shows the process of adding the code: + Code on GitHub + Work on a fork and make a pull request + Discuss changes with core developers and make improvements + Code is approved and merged Everything is in the open. We could see the code and make the changes that we needed - much better than raising a support ticket that you cannot see. We can evaluate the quality of the code and make improvements. Most code is written by a handful of developers e.g. BGS SIGMA is backed by 150+ years organisation with 600 staff, but only 2 core developers. With open source, other people can see and improve the code so it is better. The Wellerman video was released on TikTok. TikTok provides tools to allow users to "duet" and to modify videos created by others. Once you release something onto the internet, you can't predict where it will go. Our ETL Helper tool got a merge request to add a Database class. It was a great idea and something that none of us had thought of. We plan to add it into ETL Helper v2.0.

Addressing enterprise concerns

Support

Sonatype logo Anaconda logo

Security

Dependency scanning

Integration

STANDARDS!

Multiple approaches to interact with Oracle Spatial

Open source alternatives to proprietary tools

Open source in the news

Right to repair

Artificial Intelligence

An example where the corporations are flexing their muscle to ignore open source licences https://www.theregister.com/2022/11/11/githubs_copilot_opinion/ GitHub have released CoPilot. It is an artificial intelligence for code completion. It has been trained on millions of lines of code, much of which is open source (including GPL licences). Microsoft claim that the models don't contain the code itself, but some of the code that it produces is verbatim copies of open source material. The GitHub terms of service say that Microsoft can use your code. Microsoft are charging for this, but they didn't pay developers and don't generate acknowledgements. They are being sued. Since this talk was first given, ChatGPT and other AI systems have been released. Beyond licencing issues, open source is also relevant because: + Open source models will democratise AI and allow uses away from big corporations + Free software unfortunately also means that it can be used to create any specialism that you want e.g. fascist + For full transparency, it will also be necessary to publish training data as well as code.