Friday, September 28, 2007

Calling All Artists: Wikijunior Visual Math

In discussing the OLPC project, a few people at Wikibooks had shown an interest in helping, but were undecided how to do so. Using Wikijunior we could write books for the children recipients of these laptops to use for independent learning. However, as many of these children are not native english speakers, it's difficult for us to do anything directly. We could write a book with the expectation that it be translated "on the ground" by volunteers into whichever local language was necessary.

The concept also arose that we could write a book about the english language, that is a book designed to teach a non-english speaker to speak basic english without reliance on the reader's native language. Basic ideas about this involved a heavy reliance on images to associate simple words with common objects. Possibly even audio files could be added for help in pronunciation. This seems like an awfully anglo-centric solution to an unrelated problem and so this idea was scraped (at least for the OLPC project, this idea has plenty of merit among illiterate children in english-speaking countries).

From that idea we found the concept of image-based learning. We could use pictures and symbols, almost to the exclusion of all words, to teach basic subjects in a language neutral way. Using the imagemap extension we were able to come up with a proof-of-concept for the idea in [[Wikijunior:Colors]]. For the colors book, the table of contents is completely image-based: Click on a color to go to the page about that color. There are some english words in this book because it was originally designed as a simple read-along for parents and infants.

The next idea on the table was a mathematics book, which is slowly coming to fruition in the form of [[Wikijunior:Visual Math]]. This book has highlighted the need to create a large number of images representing quantities and basic arithmetic operations (addition, subtraction, counting, etc). There are some images for this purpose available on commons, but not nearly enough to complete this book. To that effect I would like to post a call to all wikimedian artists to help generate images for use in this book. Images of items in various quantities are needed most importantly (so that numbers can be directly associated with object quantities), although images that represent changes in groups according to arithmetic operations are also needed. There really is a lot of room for creativity here. I sweep through commons regularly, so images that appear in [[commons:Category:Arithmetic]], will be found and employed readily. If we can collect a large enough group of images we might even need an additional category ([[commons:Category:Numbers]] appears to already be used for a different purpose).

Thursday, September 27, 2007

Script Trading

Spent some time over on Wikinews yesterday and today, and I saw something very interesting: They have a feature where you can load a per-page javascript or css file. That means for every page there is an option to load scripts and stylesheets specific to that page only. You don't need to add bloat to MediaWiki:Common.js in order to enable features on only one or two specific pages. Wikinews hosts these files in the MediaWiki: namespace, so they are only editable by admins by default.

I've borrowed the idea, and have started a small-scale test of the script on Wikibooks. Naturally, i've changed a few details, and discussed some uses with other Wikibookians:
  1. CSS pages are loaded in per-book, not per-page. This way book authors and editors can make large changes to the appearance of a particular book by making edits to a single file (and without having to rely on volumes of specialized formatting templates).
  2. Per-book CSS files might be located in the main namespace, in the book itself. These pages wont be protected by default which means that vandals will be able to mess with them, but regular authors and editors will be able to play with them more as well. This is an idea that some other users have suggested, and is worth testing.
  3. There isn't going to be any per-book JS loaded globally, but i'm testing a way for users to do it individually. That is I can load per-book javascripts, but I can't force other users to do the same.
In return, some Wikinewsies have adapted our google analytics idea for that project. Users can opt-in to the program by installing the script into their account, and can then receive the reports by email.

Tuesday, September 25, 2007

Google Analytics, Day 1

The first results are in from my experiment with Google Analytics. As of today, I was the only user to be using this script, but more users are joining in the test. Performance issues appear to not be a factor, and looking through the generated reports, I can't find much information that would make people personally identifiable. I'll keep searching on that front though, just to make sure. If IP addresses do appear in the report that would put a hold on the whole thing (Im a CU, so I probably won't abuse that information if i do see it).

Anyway, in the first 24 hours, I:
  1. Checked my watchlist 22 times
  2. Used FireFox 100% of the time
  3. Visited from two distinct locations
  4. Had 105 recoreded page view (we had a javascript error last night, so this number should likely be higher)
  5. Viewed my monobook.js file 10 times (trying to ensure the script was installed correctly)
  6. Am located in the united states
  7. Spent an average of 1:20 viewing each page
Like i said above, last night there was a javascript error that caused my script to crash, so a lot of my hits weren't recorded. However, the numbers look to be very useful. Also, I can export the report to XML, which means that we can likely use some scripts to analyze the data and to post the results onto the wiki (if that's what people want). All in all, I would say that the test is working out very well, and I can't wait to see how things go.

Monday, September 24, 2007

Google Analytics

The issue of page counting is one that is raised regularly, not just on Wikibooks but on many other WMF projects as well. There are plenty of good reasons why the software doesnt do this automatically, not the least of which is the caching mechanism. However, in a recent wikibooks discussion, the possibility of using the free Google Analytics service has been raised. Some users, myself included, have installed the google analytics script into their own personal javascript files.

I personally access Wikibooks through multiple browsers (IE6 and IE7, Firefox, Safari), so i'll be able to test the script in a variety of ways. I'll be looking at the accuracy and relevancy of the results, as well as the performance of using this tool. If everything is on the up and up, we may start seeing a few recommendations that this script be loaded into the common.js file, at which point we will get the site-wide statistics that we've always wanted. Of course if this doesnt work out, we will go back to counting page hits the old-fashioned way: on our fingers.

Wednesday, September 19, 2007

Updated Templates

Stolen? Yes, we ripped wikipedia's new {{Ambox}} template off cold. Let's face it: it was a great idea. A complete aesthetic overhaul, and a standardization of all the various messages that we put onto a page. Plus, the nice color-coded scheme is very logical and easily usable. Red means bad, blue or green mean good. It's a system that anybody can get behind. Of course, we had to fiddle with the icons, the colors, and the CSS, but we Wikibookians can never just leave well-enough alone.

Wikibook's old template messages were ugly and archaic. When the idea was brought up to replace them en masse with a newer, sleeker, prettier and more standardized series of templates, the decision was unanimous: "yes, do it now". And so it is done. Wikibooks now has a whole set of fancy new templates, courtesy of Wikipedia.

Tuesday, September 18, 2007

Wikiquality

I got a message from Erik Moeller in my inbox today, and so should everybody else (if not you need to sign up for one of the various mailing lists). It was all about the new "Wikiquality" initiativew, involving the ne http://quality.wikimedia.org website, a new Wikiquality-l mailinglist, and the FlaggedRevs extension. It's all exciting stuff, and I think that Anthere was a genius for making reliability one of her focus points.

However, the concern that comes to my mind (as is usual) is that the whole thing seems very wikipedia-centric. Even on the quality.wikimedia.org website, it's all about wikipedia: not a single mention of any of the other sister projects. I can understand it, the whole issue was started because of reliability concerns over Wikipedia, and the negative press that it has received because of it's spam and vandalism and biased editing. I do understand that this is a good solution to a problem that Wikipedia has faced more then any other sister project.

However in this movement to improve Wikipedia I can see the catalyst for the explosive growth of other projects, specifically Wikibooks. One of the primary reasons why Wikibooks textbooks are not currently used in school class rooms, is precisely because of the non-static nature of those books. Teachers need a stable and reliable textbook to base a class around, and things would go to hell if the book changed mid-semester. Imagine if a classroom of young children reading a Wikijunior book woke up one morning to find pornographic image vandalism in their textbook? Stable versions would help to prevent that, and therefore make the whole experience more reliable and more safe.

Instead of hunting down the specific revision IDs of the pages in a book, a teacher could simply ask that the "current version" of the book be stabilized for the duration of the school period. Work could still continue in the background, and after the classes are over the new version could go live. Up till now, most classes have used Wikibooks as a collaborative writing medium, and we would really like to expand that to become a reading medium as well. I think this new quality assurance initiative is the last piece to the puzzle for Wikibooks to finally become a world-class source for reliable, high-quality textbooks.

Monday, September 17, 2007

Simple/English Wikibooks Merger

SB Johnny and myself have left a message on the [[Wikibooks:Simple Talk]] page of the simple.wikibooks project (the equivalent of the "staff lounge" or "village pump" that readers might be familiar with). The gist of the message is that we move to merge the simple.wikibooks project into the en.wikibooks project. Earlier this week, import was enabled between the two projects, which would make the transition easier then it would have been otherwise.

Initial reaction seems positive. We've tried to structure the suggestion in a way that is non-hostile. We want to accentuate the positives of a merger, without throwing around words like "failed project". Some of the positive reasons for a merger are:
  1. More administrators. En.wikibooks has more administrators, along with bcrats and CU's necessary to really keep cleanup tasks under control, and to effectively combat vandals. Simple.wikibooks currently only has two admins, one appears to be relatively inactive, and they require regular help from stewards to fight vandalism. If there is a merger, it's likely that the administrators from simple.wb could be made into admins on en.wb immediately, or through an expedited process.
  2. En.wikibooks is no longer "grumpy" (not my term for it). Our project used to copy many of the policies from wikipedia, such as the "one subject, one article" policy. We've expanded to allow multiple books on a single subject, especially if those books have different scopes, or different target audiences. This means it's perfectly acceptable to have an "X" book, along with an "X in simplified english" book. In fact, we already have a number of books written in simplified english variants, so the addition of the books from simple.wikibooks would not be any kind of problem or exception.
  3. En.wikibooks has really been expanding it's organizational structures recently. We have an entire namespace dedicated to organizing, and we've developed a number of category/DPL techniques for keeping books organized. We also have a very successful "Featured Books" program that helps to promote our best books.
  4. En.Wikibooks is very stable, with a solid group of contributors, large numbers of new accounts created regularly, new content being created regularly, etc. We are no longer suffering from the cyclic periods of ups and downs that smaller projects experience. Work is able to progress at a more even and consistent rate. This stability attracts new contributors who in turn bring more content and more visibility for our project.
There are still a few more active users on that project that we want to hear from, but overall i think it's a good idea that will bring benefit to all users. If they don't like the idea, that's fine too. I dont want to force anybody to do anything that they dont want to do, but I hope that other people can see the same kinds of benefits that I see from it.

Sunday, September 16, 2007

Adopt a Book

Books, especially younger books, need a strong driving force to keep them growing and to keep them on track. So often, the author of a new book abandons it before the book has reached a stable and mature status. Without a certain amount of stability, new editors don't contribute to the book, and it languishes over time. Wikibooks has many books that really just need to be adopted by new authors, authors who have the necessary vision and energy to turn a forgotten book-stub into a great educational resource.

[[Human Anatomy]]: Most of the links in the TOC are red, and the few blue links point to mish-mashed pages with all sorts of random content. This book has also been nominated for a merger with [[Anatomy]], which is in a similar condition. Compare this book with the fantastic [[Human Physiology]] book, which has tons of content and images.

[[Semiconductors]]: I started this book as a way to combine several stub-books and pages on this topic. Unfortunately, it isnt an area of expertise and I haven't been able to add to it much. Proof that even an experienced author can get in over his head. Since many of the pages were merged in from other books, there is very little continuity of discussion, nor commonality of formatting.

[[Geometry]]: There are many links in the TOC of this book, but most of them point to stub pages. Also, all the pages are named in a very non-descript fashion, such as "Geometry/Chapter 1". Comparse this to an excellent geometry book for younger readers at [[Geometry for Elementary School]].

[[Psychology]]: A subject that so many students study, but for which we don't have a good book. This book has only a few pages out of a relatively large outline. The few sub pages are either stubs or outlines themselves.

This is just a very small listing of books that need some help. Hopefully, i'll be able to list a few other good ones here eventually.

Monday, September 10, 2007

Dewey Decimal System

So we go through all this effort to implement the Dewey Decimal Classification (DDC) system, and now it seems the effort may have been wasted. We've created categories and templates, and all sorts of other pretty things to manage this new system, taking extra precautions to ensure that the system was both usable and expandable. Using categories and DynamicPageLists we were able to create a whole swarm of dynamically-updated lists of books by subject.

Recently en.wikibooks got a whole new namespace to use for organization, and we've been putting it to good use. For a look at the initial work (most of it bot-created) in organizing, you can travel to [[Subject:All Subjects]].

What we have now is (at least the beginnings of) a powerful new system for organizing books: logically arranged, automatically updated, completely uniform. In short, we just don't need DDC anymore, especially not with it's obvious deficiencies in the areas where Wikibooks needs the most precision (computer books).

Beyond that, some questions have come to light about the legality of using the DDC. It is under copyright, although the extent to which it is copyrighted is a matter of some ambiguity. The OCLC website has the following to offer in it's FAQ:

"The DDC is owned by OCLC Online Computer Library Center, Incorporated ("OCLC"). We do consider licensing arrangements for the DDC database. To request a licensing proposal, please send an e-mail message to DeweyLicensing@oclc.org, describing in detail your proposed use of the DDC."

Do we need a license even if we aren't using the database? Do we need a license to apply numbers to books, or do we only need a license to use the categories that OCLC has already created? If the system is a copyvio, we might have to delete it whether we want to keep it or not.

Monday, September 3, 2007

More organization, a whole namespace

Over the weekend, en.Wikibooks got a new namespace, the "Subject:" namespace. it's goal is to be an organizational namespace, where we can list and cross-list books based on a variety of metrics. By using a comprehensive set of categories, templates, along with clever use of special pages and DPL, we can automatically generate a large number of helpful book lists. The only problem so far is that there are so many possibilities, that nobody quite knows where to start.

Also, it will take much more effort to create templates and categories, and then properly tag all the books, to help maintain all the millions of lists that we are likely to develop.