Wikimedia Conference 2014/Documentation/24
24 Software development as a new opportunity for chapters
- Speakers: Emmanuel Engelhart (WMCH), Michael Maggs (WMUK)
tl;dr:
- Don't be too ambitious. Many things will fail, but that's not a problem. Involve volunteers.
- WMUK is looking at a project to automatically rate article and media file quality.
Notes
[edit]Emmanuel Engelhart (WMCH) soft dev.
[edit]Introduced his background, here makes the speech as a developer.
- Why it is important to invest in software? The question of working with free knowledge is how we have to do it.
- It is not good to let WMF decide what is necessary.
- Software projects are risky, it is possible to lose much money.
- There are 2 types of projects: integration or software dev (from scratch). Focus is on the last.
- It must be ae an emancipation factor
- Free software is a good decision because of access to the source code, save privacy
- It is important to deal with the endusers
- No solution works out of the box. If you by an external solution, please test it.
- Thoughts about Software development projects
- Chapters want to be big players in software projects but don't have money
- there are many volunteers with computing background
- you have to work with them as partners
- KIWIX is a volunteer project (7 ways to go offline); see http://www.kiwix.org/wiki/Main_Page/de
- WMCH is not big enough, but there is much to do in development
- KIWIX is used in Afripedia and other projects, also with hardware
- volunteers work for fun - make that happen
- more people are working for money - that is not the same
- support people with information and publishing - that costs not much and is important for them
- connect these volunteers with people in international teams (MediaWiki, Softewaredev come together, ...)
- many of these vol developers have no project management - support this for them, Chapters and WMF have the possibility
- you don't need much money
- important is a project page, takes one hour but is very important
- GLAMWikiToolSet
- plan to work with WMF
- involving at the beginning is important
- see Commons:GLAMToolset_project
- Conclusion
- not be to ambitious
- much will fail - that is not a problem
- give projects a change
- most important are the people, right people can do enough
- it works not like Wikipedia does
- Try to involve volunteers!
Michael Maggs (WMUK) - WikiRate: rating Wikimedia
[edit]
A plan for tech work that is being considered by WMUK. How to rate article and media file quality.
For more details, see https://wikimedia.org.uk/wiki/Technology_Committee/Project_requests/WikiRate_-_rating_Wikimedia
How can you measure your impact? Target must be smart. This is an overview, what can be done.
- Increasing need for chapters to measure using SMART targets
- SMART means Specific, Measurable, Attainable, Relevant and Timely
- Measuring article and media file quality (as opposed to quantity) is very difficult and expensive at the moment. Mostly done by hand.
- We want to provide an easy-to-use set of tools to rate article and media file quality.
- Other Work
- WikiMetrics https://metrics.wmflabs.org/
- WikiTrust
- wm_metrics by WMFR
- SuggestBot
- ClueBot
- Who could use?
- Chapters to measure their impacts
- Educationalists
- Individual editors
- Academic and Wiki researchers
- Foundations analytics team
- How to create the model
- Score some articles/media files manually
- List of possible inputs, and desired outputs
- Adjust (train) the model using standard machine learning techniques
- Apply the trained model
- Use trained model to rate articles and media files automatically by quality.
- Keeps FDC happy if you can report hard numbers to them!
Q&A
[edit]- Use the Mailinglist List from the foundation to find projects and other developers
- How to extimate Softwareprojects?