A Tale of an Encyclopedia in Graphs

Just one more post about Voxopedia.

Vox Day has an additional blog as a kind of support site for his Wikipedia knock-off. https://infogalactic.blogspot.com.au/2016/10/500-editors.html

We now have 500 registered Galaxians and we’re seeing a growing number of daily edits as well.

We’ll be introducing three levels of editing capabilities soon.

L1: Create and edit new pages only.
L2: Create and edit new pages, add text to existing pages only.
L3: Create and edit new pages, add and delete text from existing pages.

Once we get to Phase Three, this sort of differentiation will be unnecessary, but it will reduce the likelihood and amount of edit-warring in the meantime.

It is true that the number of registered ‘Galaxians’ is over 500 but it is very wrong to see that figure as having much relevance to how the wiki is performing. It may be that 500+ is a goodish estimate of the wiki’s supporters but it has very little relation to the editing activity on the site.

I’ve pulled off the Recent Changes page, a data set that looks at edits to main pages (i.e. not Talk: pages, User: or other special pages) that are not flagged as ‘minor’ edits, that dates back to 4 October. From that data set you can see how much editing is being done and by whom (editors are pseudonymous and the data is public so no privacy issues are in play).

This is how the edits are playing out cumulatively:

[more below]

It isn’t quite linear but it is not a increasing rate of daily edits. There are splurges of activity but they themselves are revealing. For example, there is a jump in the number of edits that occurs during 22 October.  The bulk of those edits is by one editor, running through a large number of pages changing ‘pro-choice’ to ‘abortion legalization advocacy’ in category headings on the grounds of ‘Removing bias from category title’. Those 68 changes represent about 6% of the edits in this data set.

Other changes in the data set that might count as ‘minor’ but weren’t flagged as such including fixing redirects and fixing links to the real Wikipedia that had been included in the original Wikipedia pages as proper URLs instead of Wiki codes.

Substantive edits are much rarer. Kudos though to ‘Bassiano’, a prolific editor who has been systematically adding public domain pictures of Victorian/Edwardian artists & poets – I’m not being sarcastic, that is the sort of task that helps add value to a wiki. However, that work isn’t doing much for Voxopedia’s core problem.

Put simply it isn’t growing fast enough. Once the files were copied from Wikipedia, those pages became out of date. Any page that relates to changing events is stuck in the past (early 2016?) on Voxopedia unless somebody updates it. On top of that, Wikipedia is spawning new pages and new content. In addition to both of those, Voxopedia is supposed to cover topics and issues that the real Wikipedia is supposedly censoring etc etc.

New pages, for example, are being added very slowly and as I’ve pointed out, many are of poor quality. Here is a graph of new pages in the main space (i.e. not User pages etc).

 

That is so linear that I had to draw a line through it.

It is pretty much straight because it mainly shows the number of new pages that a small distinct group of people can create. Adding more members isn’t impacting on the number of new pages being added because the new members aren’t doing anything.

The problem with becomes clearer when looking at the proportion of edits per person.

Two people alone account for nearly 70% of all the edits in the data set. In terms of ALL edits then a third person, ‘Crew’, adds a lot more as they primarily deal with new members.

There is a long tale of names in what looks like a Zipf-like distribution. In total 73 names have made edits in this data set – approximately 15% of the number of registered users.

In short, Voxopedia changes slowly and by a small number of people.