Thursday, 22 January 2009

Small screen$

Correction

Angela subsequently provided a correction for this post; the error was in the cited publication:
"The new platform has nothing to do with Nokia as far as I know. I've sent a correction to the author of that article. The Foundation already has a branding deal with Nokia but that's not related to this."

Angela Beesley (former member of the Wikimedia Foundation Board of Trustees, and current chair of the Advisory Board), has indicated during a talk at Linux.conf.au that the Foundation will be announcing a new mobile platform for Wikipedia later in the year. According to ZDNet Australia, the platform is currently under development and will be licensed to Nokia.

There are already a number of iPhone apps for reading Wikipedia, whether online and offline, including plenty of good free apps. In addition to dedicated apps, there is a solid specialised Wikipedia mobile interface, and there have been efforts to make the regular web interface of Wikimedia wikis more practicable in mobile browsers.

It's good news, however, to hear that a dedicated Wikipedia interface is on its way for one of the closed mobile platforms. What's also interesting is to hear that the platform will be licensed to Nokia, which makes it sound as if there's some commercial arrangement involved. It's interesting in that Angela's talk also touched on the recent success of the fundraiser, and the possibility of alternative sources of funds, such as selling physical versions of content like books and posters.

Licensing a branded mobile platform strikes me as an interesting potential revenue stream. It reminds me of Mozilla's arrangement with Google: something that benefits users, and also allows money to be made without compromising on principles (unlike, say, introducing ads).

Tuesday, 25 November 2008

The human touch

Google has recently released SearchWiki, a set of tools for annotating Google search results. It's a rather dramatic change to the main search page, accesible by anyone logged into a Google account.

The interesting thing is that it seems like a personal version of Wikia Search (personal in the sense that only your alterations change the order of results, although you can see comments from everyone), though earlier comparisons were made more to link sharing sites like Digg.

So, might the emergence of this tool mean that Google is not, unlike so many others, underestimating the potential of Wikia Search?

Tuesday, 18 November 2008

Bug statistics

Since the beginning of September, the bug tracker for MediaWiki has been sending weekly updates to the Wikitech-l mailing list, with stats on how many bugs were opened and resolved, the type of resolution, and the top five resolvers for that week. With eleven weeks of data so far, some observations can be made.

The following graph shows the number of new, resolved, reopened and assigned bugs per week (dates given are the starting date for the week). The total number of bugs open that week is shown in blue, and uses the scale to the right of the graph:


The total number of open bugs has been trending upwards, but only marginally, over the past couple of months. It will be interesting to see, with further weekly data, where this trend goes.

It also seems that the number of bugs resolved in any given week tends to go up and down in tandem with the number of new bugs reported in that week. Although there is no data currently available on how quickly bugs are resolved, I would speculate that most of the "urgent" bugs are resolved within the week that they are reported, which would explain the correlation.

Note also the spike in activity in the week beginning 6th October; this was probably the result of the first Bug Monday.

The second graph shows the breakdown of types of bug resolutions:


The distribution seems fairly similar week on week, with most resolutions being fixes. It's interesting to note that regularly around 25% to 35% of bug reports are problematic in some way, whether duplicates or bugs that cannot be reproduced by testers.

The weekly reports are just a taste of the information available about current bugs; see the reports and charts page for much more statistic-y goodness. And kudos to the developers who steadily work away each week to handle bugs!

Sunday, 31 August 2008

How Collective Wisdom Shapes Business, Economies, Societies and Nations (and Wikipedia articles)

Alaska Governor Sarah Palin was selected as presumptive Republican presidential candidate John McCain's running mate on Friday, and her Wikipedia article has seen a predictable explosion in editing activity. From the article's creation in 2005, up until the announcement on Friday, the article had been edited something like 900 times. Since then, however, it's been edited nearly 2000 times again.

What's more interesting is how the article was edited before the announcement was made. Ben Yates mentions this NPR story detailing edits made to the page by a user called Young Trigg, who may or may not have been Palin herself (or someone on her staff). But Young Trigg was not the only person editing the article.

The Washington Post reports on some analysis done by "Internet monitoring" company Cyveillance, which found that Palin's article was edited more heavily in the days leading up to the announcement than any of the articles on the other prospects for the nomination. A similar pattern emerged in relation to the articles on the frontrunners for the Democratic vice-presidential nomination: Joe Biden's article was edited more heavily than the other potential picks in the leadup to his selection as Obama's running mate last week.

Also similar were the types of edits being made: both Palin and Biden's articles saw many footnoting and other accuracy-type edits in the leadup to the announcements of their selection. As a final piece of intrigue, the editors making these edits about Palin and Biden were far more likely to also be actively editing McCain and Obama's articles respectively than were the editors editing articles on the other potential nominees.

There are at least two explanations for these patterns. The first is that the two campaigns, knowing full well who the nominees would be, were editing the articles in advance of the announcement to ensure that they were accurate (or to take the cynical view, to ensure that they were favourable), knowing full well that Wikipedia would be one of the major sources of information for the public - and for journalists and campaign staff too - following the announcements.

The alternative is more interesting, to my mind. Cyveillance, who did the analysis, is usually in the business of data mining in the business world, aiming to collate disparate sources of public information to predict financial and commercial events before they are publicly announced. Wikipedia may be performing exactly the same function: a variety of editors collating disparate pieces of information in a far more powerful way than any individual could. It's already (un)conventional wisdom that the betting markets are equal or better predictors of elections than opinion polls are: a basic application of the efficient market hypothesis. In a similar way, high profile, highly edited Wikipedia articles like these are the marketplace of the information economy.

Saturday, 23 August 2008

Userpage Google envy

Brianna Goldberg, a Canadian journalist with the National Post, wrote on Friday about her efforts to become the number one Google result for her name. Her quest was sparked by discovering that the Wikipedia user page of another Brianna Goldberg was ensconced in the top spot.

The journalist Goldberg obtained advice from search engine optimization experts on methods for advancing her ranking, but still had difficulty displacing the userpage. Moreover, the article on the journalist comes in second to the userpage in results from Wikipedia. Wikipedia user pages are certainly highly visible: every time you sign an edit, you're creating a link to your user page.

This relates to a discussion from last month on the mailing list about whether user pages (and certain other types of pages) should be indexed by search engines at all. The Wikimedia sites already instruct search engines not to crawl certain pages, including deletion debates, requests for arbitration pages and requests for adminship, but there have regularly been calls for more types of pages to be restricted (see here for example).

So, should user pages be blocked from search engine crawlers?

Monday, 18 August 2008

US court groks free content licensing

The US Court of Appeals for the Federal Circuit handed down an interesting and significant decision on Wednesday, which could have a number of valuable implications for the validity of free content licences.

The case, Jacobsen v Katzer, was about software for interfacing with model trains. Robert Jacobsen is the leader of the Java Model Railroad Interface project (JMRI), which releases its work under the Artistic License 1.0; Matthew Katzer (and his company Kamind Associates) produce commercial model train software products. It was alleged that either Katzer or another employee of Kamind took parts of the JMRI code and incorporated it into its own software, without identifying the original authors of the code, including the original copyright notices, identifying the JMRI project as the source of the code, or indicating how it had modified the original JMRI code.

Jacobsen sought an interlocutory injunction, arguing that since Katzer and Kamind had breached the Artistic License, their use of the JMRI code constituted copyright infringement. However, the District Court considered that Jacobsen only had a cause of action for breach of contract, not for copyright infringement, and because of this Jacobsen could not satisfy the irreparable harm test (in the case of copyright infringement, irreparable harm is presumed in the 9th Circuit), and was not entitled to an injunction.

Jacobsen's appeal to the Court of Appeals was against this preliminary finding. An assortment of free content bodies (including Creative Commons and the Wikimedia Foundation) appeared as amici curiae in the case, submitting an interesting brief containing a number of arguments that the Court of Appeals seemed to agree with.

The legal issue at stake in the appeal concerned the difference between conditions of a contract and ordinary promises (covenants, in US parlance). If a term in a contract is a condition, then the promisee has a right to terminate the contract. In the context of a copyright licence, if someone using the licensed material breaches a condition of the licence, they are then open to a copyright infringement action (unless they have some other legal basis for using the material). Contract law will still hold someone responsible for breaching a contractual promise, but the remedies are different, and as was the issue here, it's much harder to get an interlocutory injunction.

Whether or not a term is a condition is a matter of construction, and depends on the intention of the parties. In answering the question of whether the relevant terms were conditions, the Court of Appeals made a number of important observations which are applicable to free content licences generally.

The first observation was that, just because with free content licensing there is no money changing hands, it is not the case that there can be no economic consideration involved. The Court recognised several other forms of economic benefit which free content licensors derive from licensing their works:

"There are substantial benefits, including economic benefits, to the creation and distribution of copyrighted works under public licenses that range far beyond traditional license royalties. For example, program creators may generate market share for their programs by providing certain components free of charge. Similarly, a programmer or company may increase its national or international reputation by incubating open source projects."

This is a really significant observation for the court to make, because there are some major ideological barriers that seemed to get in the way of the District Court on this point. Even though free content licencing is all about authors dealing with their economic rights under copyright, free content is all too often viewed as non-economic. Just because free content doesn't fit in with the traditional royalties-based system, it does not mean that there are not real economic motives involved.

The second observation was made in the context of the general rule (applicable in that jurisdiction) that an author who grants a non-exclusive licence effectively waives their right to sue for copyright infringement. If the relevant terms were conditions, then they would be capable of serving as limitations on the scope of the licence, which would negate this rule. The Court said that:
"[t]he choice to exact consideration in the form of compliance with the open source requirements of disclosure and explanation of changes, rather than as a dollar-denominated fee, is entitled to no less legal recognition."

Again, this seems to be an important point in terms of getting over psychological hurdles. The District Court was clearly hung up on the terms in the Artistic License allowing users to freely distribute and modify licensed material; it focused on the breadth of the freedoms granted. In doing so it overlooked that while the License did grant broad freedoms, it clearly circumscribed them. The Court of Appeals understood what the District Court did not: that releasing material under a free licence is not the same as giving it away.

The heart of the decision was of course about the particular wording in the Artistic License. The use of the phrase "provided that" in the Artistic License was significant, because such wording usually indicates a condition under Californian contract law. Further, the requirement that any copies distributed be accompanied by the original copyright notice - a relatively common term - also typically indicates a condition.

In the end, the Court of Appeals decided that the relevant terms were conditions, and that Jacobsen had a copyright infringement action open to him. Since the District Court didn't assess Jacobsen's prospects of success on the merits, the Court of Appeals remanded the injunction application back to them for their consideration. Given that Katzer and Kamind apparently conceded that they did not comply with the Artistic License, Jacobsen would seem a good chance to get his injunction, and later to succeed at the merits stage.

Though much turned on the particular wording here, the reasoning behind the assessment of the terms can easily be applied to other free content licences, as can the recognition of the economic motives involved in free content licencing, motives which though non-traditional, are both legitimate and worthy of protection by the law. Independent of any value as a binding precedent, this case is a magnificent example of a court really appreciating the vibe of free content.

Monday, 28 July 2008

Kno contest

Google's Knol was opened to the public last week, to much fanfare. When Knol was announced in December last year, it was immediately compared to Wikipedia, and the comparisons keep coming now that it has launched. However, as I wrote at the time, the comparison seemed to be wide of the mark in many important ways. Now that Knol has launched and we can see how it will actually work, I think the accuracy of the comparison is still not borne out.

The three key differences I noted at the time were the lack of collaboration in writing knols, the plurality of knols (more than one on the same subject) and that knols will not necessarily be free content, differences which go to the core of what makes Wikipedia what it is.

As it turns out, Knol does provide a couple of options for collaboration, allowing authors to moderate contributions from the public, or allow public contributions to go live immediately, wiki-style. The other mode is closed collaboration, but it does allow for multiple authors at the invitation of the original author.

As the sample knol hinted, Knol does provide for knols to be licensed under the CC-BY 3.0 licence by default, and allows authors to choose the CC-BY-NC 3.0 licence, or to reserve all rights to the content. However, these are the only licences available; in particular, no copyleft licences are available.

Of course, the thing to remember is that Knol is an author-oriented service, so even if an author selects open collaboration and the CC-BY licence, it appears that they can change their minds at any time, and, for example, close collaboration on a previously open knol (I might need to do some closer reading of the terms of service, but it would also appear possible to revert the Knol-published version to all rights reserved model, too).

The author-oriented approach is apparent in most of the features of Knol. On a knol's page you don't see links to similar knols, or knols on related topics (as you would on a Wikipedia article) you see links to knols written by the same author. Knols aren't arranged with any kind of information structure like Wikipedia categories, or even tags; the URLs are hierarchical, but there knols are gathered under the author's name.

No, Knol is not a competitor to Wikipedia (or at least, it's competing for a different market segment). It's more a companion to another Google property, Blogger. It's a publishing platform, but not for diary-style, in-the-moment transient posts; it's for posts that are meant to be a little more timeless, one-off affairs. Google say so at their Blogger Buzz blog:

"Blogs are great for quickly and easily getting your latest writing out to your readers, while knols are better for when you want to write an authoritative article on a single topic. The tone is more formal, and, while it's easy to update the content and keep it fresh, knols aren't designed for continuously posting new content or threading. Know how to fix a leaky toilet, but don't want to write a blog about fixing up your house? In that case, Knol is for you."


Some of the content on Knol might start off looking like Wikipedia articles, but over time I'll bet that the average "tone" of knols will find a middle ground between blogs and Wikipedia's "encyclopaedic" tone as people come to use Knol as a companion to blogging.