Cory Doctorow wrote an interesting little column for Make magazine about Wikipedia and the way that it “contains facts about facts” [1]. One of the issues with this is that you can’t (or at least shouldn’t) make corrections based on your own knowledge of a subject, you need to cite references. Cory gives a hypothetical example where a world renowned expert couldn’t simply fix an error in Wikipedia, but they could give an interview with the New York Times decrying the quality of Wikipedia and citing the correct data – then once the NYT article was published it could be referenced as an authoritative source and the page updated.
Mako wrote about the Wikireader (an offline reader for Wikipedia) and advocated a facility for writing changes while offline and uploading them [2]. From the perspective of improving the state of the art in computer science and making new features available for more users this sounds like a great idea. But in terms of fitting in with the way Wikipedia works it doesn’t seem viable to me.
Any time that you need to use an offline Wikipedia reader it seems that it will either be impossible or unreasonably difficult to gain the net access necessary to find the reference that Wikipedia requires. Sure you could fix spelling errors, reformat paragraphs, etc but not do anything serious. I wonder how many people actually put any serious effort into Wikipedia with the sole purpose of making it more readable. In terms of my own editing I have made a number of changes that improve readability, but all of them were due to reviewing pages for technical accuracy.
One of the comments on Mako’s post suggested that instead of having an option to directly edit pages there be an option to take notes on pages which could then be used for later editing. Maybe something like a two button combination to flag a particular paragraph as being of dubious accuracy so that the user could later do some Google searches for references.
Of course another problem with this is the delay in updating a Wikireader. For most reading of Wikipedia it doesn’t matter if the content is 6 months old, so we should expect that most users of the Wikireader don’t have the latest data. The Wikireader company offers an update service of posting a new memory card twice a year [3]. Given postage delays etc we should expect that the average age of the content is about 4 months for people who subscribe to the update service, and probably a lot greater for people who download the updates. The popular pages of Wikipedia change fast, in the space of 4 months a page can change significantly.
In terms of general Wikipedia operation I wonder if it would be beneficial to have a button to flag content as possibly not meeting the Wikipedia guidelines. It seems that the number of people who can recognise low quality pages is going to be significantly higher than the number of people who have the skill and time to fix them. I’m sure that there are many people who would just love to be able to choose from a list of Wikipedia pages that have received many negative votes. Such votes could be stored offline and uploaded later, but again that would rely on having recent content on the reader (it would be annoying for a recently fixed page to keep getting voted down).
One area where offline editing of a Wiki would work well is that of niche Wikis. For example if I downloaded a copy of the Debian Wiki (which would only take a tiny fraction of the space in a Wikireader) then I could easily update it every week and the incidence of other people editing a page in the mean time would often be low. I could write long updates that make significant changes to SE Linux related pages while offline without needing any references and in many cases without much risk of conflicts. Implementing a distributed version control system to manage such updates shouldn’t be difficult in principle, although the VCS might have to run on a laptop or desktop system after taking raw data from the Wikireader (tools like GIT seem memory hungry and might not fit on a small device).
Also Mako refers to the “already indefensibly large gap between the number of readers and editors on Wikipedia“. Like Mako I don’t have a TV watching personality. I agree that it is desirable to facilitate read-write access for everyone, but that doesn’t have to involve write access to Wikipedia. I am not convinced that there is a difference between the number of readers and editors of Wikipedia that indicates a problem here. As something becomes more popular the number of people who use it without being committed to it increases, and it also becomes more integrated into society. I don’t think it’s a problem that many people use their writing time for other tasks such as blogging rather than editing Wikipedia. Wikipedia makes a great reference for blog posts and blog posts drive readers (and thus potential editors) to Wikipedia.
Finally I think we should consider the fact that different people have different skill sets. Maybe the Wikipedia reader base gained a significant portion of the people who have the skills to make good editors a long time ago and now the majority of new users just don’t have the skills. If that is the case then having the ratio of readers to editors change in favor of readers while the number of readers expands would be a good thing. I don’t think that I am being elitist in acknowledging that different people have different skill sets, and that some portion of the population will lack the skills necessary to make good contributions to Wikipedia (good enough to outweigh the mistakes that they make). I think that valuing the contributions of all people in society does not require that we value all contributions to Wikipedia.
