This page lists interesting things that might inform design decisions made on Civilization.
Contents
Board Listings
Inspirations
Board listing with recent threads included (archived)
Very solid reskin of VBulletin. (archived)
WoW has some very interesting minimalist forum design (archived)
Damnations
Font size way too small even though there's tons of open space, kind of bland. (archived)
Predictions
Inspirations
Good Judgment Open
In many ways Good Judgment Open (archived) is the bar in group prediction software. This is because it's at the heart of the research undertaken by Philip Tetlock to make predicting the future more rigorous. It has institutional backing and a wide variety of participants posing questions to the crowd. If you can't meet or exceed Good Judgment Project, you're probably wasting your time.
What do they do right?
-
Provides a training program explaining the basics of sane prediction.
- Well defined scoring mechanism which is explained to the user during training.
-
Button to 'request resolution', which to my memory Metaculus did not have.
- Tagging exists, and is present as an option to search by in the regular search interface.
-
Powerful search interface for finding questions exists.
- Interface allows user to 'follow' a question.
-
Interface is sleek and modern looking, professional.
What could use improvement?
-
Knowledge aggregation seems weak. Comment sections exist on questions, but no comprehensive way of putting together the communities knowledge on a subject into a single artifact. A wiki page associated with each question seems like it could be a huge help here. This problem becomes even more important when you consider similar questions may be asked again in the future and previous research could save time.
-
Seemingly limited views on forecasts of others. For example there's a general consensus but no option to look at how people with high brier scores predicted as compared to the average. (Perhaps this is a feature, not a bug since it stops people from just following the leader?) No more detailed views seem to be available for closed questions.
- Training course is done through godawful Flash based interface. Could definitely use upgrade to HTML5/CSS3/JS.
- Avatars done through gravatar, which is awful.
Metaculus
Metaculus by contrast to PredictionBook appears to be making a serious attempt at things (archived).
Lets talk about what they do right:
-
Multiple views on the questions, so that I can sort by the exact basic categories I outlined as lacking on PredictionBook.
-
Every question appears to have explicit terms detailing what would constitute a positive resolution of the question. PredictionBook by contrast often tolerates lower quality questions which may even have no description at all.
-
The questions are indexed by category, so you can look at specific fields and areas of expertise to make your predictions.
- Public Leaderboard, a feature which seems incredibly basic that PredictionBook manages not to have.
-
The interface is interesting and even the layout of individual predictions is food for thought on how Civilization should handle the prospect.
Damnations
PredictionBook
PredictionBook is actually quite mediocre (archived)
Lets count the ways:
- Exactly one view on the data with two graphs (one not pictured in archival) (archived) this means that as prestige building tool PredictionBook is of limited value since it's hard to prove any of those numbers really mean anything.
- Also exactly one view on the predictions made on predictionbook (archived), worse still it's the most recent predictions, which thanks to the nature of predictionbook are often intensely personal garbage of no interest to anyone other than the predictor. It wouldn't be possible to break this up into a "most viewed/hottest predictions", "most responded predictions" sort of deal? Ludicrous, this is pure lazy programming or incredibly incompetent database design on the part of the site owners.
- Totally spammed up with crappy no-signal predictions of a personal nature (archived) as described above.
- Because of limited data analysis and public nature of predictions (haven't actually signed up yet, but I probably should just to completely evaluate) it's not even very good as a personal prediction journal.
- Garish interface and layout, it's kind of petty but it needs to be said.
Export
Inspirations
Honestly we could sit here all day and recount data export horror stories, but all you really need to do is follow Twitter's example. Everything Twitter does when it comes to data export is Correct and for this editor they raised the standards of what to expect from an export. Here are the relevant features:
- Have an explicit export feature which exports useful data from ones profile.
- Export data in an easily machine readable format.
- In Twitter's case, export it in multiple easily machine readable formats (here JSON and CSV).
- In the export, include a reader for the data so that someone doesn't need programming ability to make use of their export.
- Make that reader quite good.
Pinboard
The Pinboard (archived) export UI couldn't be simpler:
Lesson:
Simple is often best. At the very least, start with the simple way; provide your users with the simple, obvious, basic way of doing a thing, and in the meantime you can work on developing some complex, fancy, advanced way of doing it. (But remember: by the time you get it done—if ever—you might find that your users like the simple way. If so, be sure to leave the simple way, and add the fancy way as an alternative!)
Melding Wiki + Forum
Inspirations
ENWorld
ENWorld (archived) is a very popular (and rather old, having been founded in 1999) tabletop (pen & paper) roleplaying game discussion forum and community. It runs on (what appears to be a customized version of) VBulletin.
Wiki Threads are a feature added this year.
“How to use a WIKI THREAD” (archived) on ENWorld
List of wiki threads (archived) on ENWorld
TODO: Analyze the features, implementation, and UX of ENWorld's wiki threads feature
