Search Wikia, the open source search engine from Wikipedia co-founder Jimmy Wales’ Wikia, Inc. organization, launched today in an alpha version. Widely hyped, typically portrayed as a heroic David vs. Goliath struggle, it’s really just yet another crappy search service that may, potentially, if all goes well, eventually turn into something useful.
For now, though, Search Wikia is essentially useless as a search engine. What makes it interesting are some of the ideas and approaches that the service is exploring. I’ll not bother bashing the search results that even Search Wikia itself admits are poor. Instead, I’ll take a look at some of the problems and issues with the current state of web search today that Search Wikia is attempting to address.
“We’re not producing a Google killing search engine”
That quote above is from Search Wikia founder Jimmy Wales himself in a Q&A with Jimmy Wales a little over a year ago. In the Q&A, Danny probed Wales for his motivations for creating a new search engine. Wales singled out several issues, which have since evolved into “organizing principles” for Search Wikia (more on those below).
One problem, according to Wales, is that search engines are effectively black boxes—you punch in a few keywords, the search engine does its thing, and serves up results without any explanation. The rub is that you have no idea what just happened, and sometimes search fails miserably.
“For certain types of searches, search engines are very good,” said Wales. “But I still see major failures, where they aren’t delivering useful results. I think at a deeper almost political level, I think it’s important that we as a global society have some transparency in search. What are the algorithms involved? What are the reasons why one site comes up over another one?”
In a New York Times article today, Wales said, “I think it is unhealthy for the citizens of the world that so much of our information is controlled by such a small number of players, behind closed doors… We really have no ability to understand and influence that process.”
Which leads to the next problem: Lack of community involvement. This year, we’ve seen an explosion in the popularity of social media sites, including efforts by the major search engines to include people in the mix, in many ways. But not to the extent that Wales feels is appropriate.
How has Search Wikia addressed these problems? Let’s take a closer look at Search Wikia’s “four organizing principles” featured prominently on the home page of the Search Wikia Labs Wiki.
Four Organizing Principles (TCQP) – the future of Internet Search must be based on:
- 1. Transparency – Openness in how the systems and algorithms operate, both in the form of open source licenses and open content + APIs.
- 2. Community – Everyone is able to contribute in some way (as individuals or entire organizations), strong social and community focus.
- 3. Quality – Significantly improve the relevancy and accuracy of search results and the searching experience.
- 4. Privacy – Must be protected, do not store or transmit any identifying data.
Let’s look at each of these four principles and how Search Wikia is dealing with them in more detail.
Transparency: Fostering democracy, or handing the fox keys to the henhouse?
Search Wikia is an “open source” search engine, meaning its users are free to modify and adapt the software, search results, and other content compiled by the search engine.
In the software world, the open source movement has aroused near religious fervor in some. The idea that software source code should be freely available for anyone to view, modify, adapt, and share it with others has merit, as evidenced by the success of the Linux operating system. But open source hasn’t really displaced so-called “closed-source” software in any meaningful way. Microsoft is still by far the dominant operating system provider, despite major efforts by the open source Linux community.
And while open source efforts may work well for something like an operating system, where the software must execute clearly defined functions with outcomes that are pretty much universally agreed-upon, search results, by nature, are highly subjective, with few standard techniques or methodologies apart from basic matching, sorting, and other relatively straightforward functions that are just the building blocks of a search engine. And in the end, one person’s perfect result may be complete garbage for someone else.
Further, search engines are such complex “entities” that I’d argue there is no single individual within Google, Yahoo, or any other major search engine that actually understands everything about how the search engine works. Computer code is highly technical, and while a programmer can often get a sense of what code does by reading it, code must be run and tested along with all of the other interrelated elements to truly see how it works. I have serious doubts that a volunteer corps could make “improvements” to search engine code without introducing unintended consequences, or worse, intentional biases that may seem perfectly reasonable to the volunteer coder yet totally unacceptable to the programming community—let alone searchers who don’t participate in the process.
It’s somewhat easier with text: The Wikipedia community does a good job of monitoring and fixing incorrect entries in its encyclopedia, but how could a programming community do a similar job of policing given the highly subjective and extremely technical nature of coding for a search engine?
And as searchers, do we really want or need that transparency? Ten years ago I could look under the hood of my car and fiddle with my engine when I wanted to modify something. Today, just about every system in my car is computerized, completely inaccessible to my tinkering. But given the virtually maintenance-free operation of my car, I’m perfectly happy with that change and don’t long for the lost days of “engine transparency” at all.
In fact, though I’ve studied computer science at the graduate level, I have a difficult time deciphering the apparent “relevance score” that appears next to each Search Wikia result. Click on this link and you’ll see a page that shows you how the open source Nutch search engine used by Search Wikia went about calculating the relevance of a particular result. “Transparent,” yes, but also totally opaque to anyone other than a programmer who’s familiar with Nutch.
And what about spammers? I know several search marketers who lean toward the dark side who would absolutely love to see exactly how search algorithms work. I’d be willing to bet a sizable sum that they’d be able to figure out a way to game the system to their advantage in less than 24 hours.
In fact, it has been less than 24 hours since the launch of Search Wikia, and spammers have already discovered this new resource: Check out the Recent wiki changes page to see what various enterprising “community members” are adding to Search Wikia.
Community: But we already have community!
All of the major search engines save Ask.com have launched numerous “community” features over the past several years. Yahoo is by far the most aggressive in rolling out community-influenced services, such as Answers, My Web, and others. Google and Microsoft have their own initiatives.
Further, most have offered technically-savvy users the ability to both use and create things like widgets, APIs, and other extensions of the search engines’ functionality. These initiatives offer both community, and, to a certain degree, transparency, for anyone with the skills and interest to avail themselves. Granted, none are playing the community gambit to the extent that Search Wikia is, but it’s not clear that community in and of itself is a beneficial thing for search services. Virtually all “social search” services introduced since the beginning of the web have either failed or not garnered very much traction. This doesn’t mean that community isn’t useful—just that it isn’t going to be a silver bullet for improving the quality of search or attracting users to Search Wikia.
You can participate in the (non-programmer) Search Wikia community in many ways. You can post bug reports, suggest sites for a “whitelist” of sites considered “must have” authorities, or create a “mini page” (essentially a Wikipedia-like entry summarizing a topic).
Search Wikia is also a social network of sorts—when you register, you are given a profile page that lets you add information about yourself, which then in turn becomes part of search results if people search on your name.
When you register for a Search Wikia account, you can register both a user name and your real name. Somewhat bizarrely, though, you see this message next to the real name field: “Real name is optional and, if you choose to provide it, will be used for giving you attribution for your work. This field must be filled in to use Search Wikia.” So, your real name is optional but you must fill it in if you want to use the service. Right.
A huge problem with community-driven sites is politicization: A certain group of users invariably comes to dominate discussion and exerts undue influence over less “powerful” users. We have seen this for years in the Open Directory project; more recently in Digg and even Wikipedia itself. What starts as a noble experiment in participation typically degrades into a species of the tragedy of the commons.
Quality: Is it tuh-may-toe or toh-mah-toh?
To me, improving quality is Search Wikia’s most laudable goal—and one that it is far, far away from achieving. Sure, search quality at the major search engines is OK these days, but we’re still a long way from being perfect, and anyone you talk to from the major engines who’s honest admits this. But quality is such a vague concept, with relevant results so much of an “eye of the beholder” issue that, while notable, “improving quality” is also virtually a meaningless, almost gratuitous goal. Turn this principle around: Does anyone set out to build a search engine with poor quality? Quality is the price of admission these days, not a differentiating feature. If you don’t have good quality from the outset, you’re dead—and it may have been a mistake for Search Wikia to open itself up at this early stage when its search results are just awful.
Search Wikia has an advantage in that the thousands of Wikipedia users may opt to chip in to help improve the quality and fill in the missing blanks in Search Wikia. But this will need to happen quickly for Search Wikia to gain traction. Wikipedia survived and ultimately thrived because it filled a need (most online encyclopedias were fee-based until Wikipedia). Unlike Wikipedia, Search Wikia isn’t filling any unmet need at this point, and it’s just not clear that volunteering will ultimately reward users.
Privacy: Don’t ask, don’t tell
Search privacy has become a hot issue once again over the past year, with most of the major engines announcing new, stronger policies regarding the data they capture and store. Search Wikia says it is going farther than most of the majors to protect privacy, but there are some blazing red flags with this initial version.
Also disturbing is the broken “Terms of Service” link that appears at the bottom of each page, which currently simply just takes you to the top of whatever page you’re viewing.
In making privacy a core principle, I would have expected a lot more disclosure, transparency, and reassurance. Instead, you have to take Search Wikia’s commitment to privacy pretty much completely on faith at this point.
The Bottom Line
If you’re looking for a worthy alternative search engine, don’t bother with Search Wikia at this point—its results will make you long for AltaVista circa 1994.
However, if you support the principles that Search Wikia is pursuing and you feel like volunteering for a pretty ambitious experiment, definitely register and give Search Wikia a try. Be prepared to spend a fair amount of time learning how to use the system—and in dealing with other users who may not agree with you or don’t like your work.
If Search Wikia ultimately does gain traction and can live up to its guiding principles, we may yet see the emergence of a truly useful alternative search service. Just don’t hold your breath for that to happen any time soon.