An up-to-date look at free software and its makers

Projects on the Move

By Carsten Schnober and Martin Loschwitz

Ever since the Wikipedia free encyclopedia took off, creator Jimmy Wales has been on the lookout for new projects. Wales' latest idea is a free, community-driven search engine that is inspired by Wikis, which he hopes will compete with the the global players like Google.

Next Generation

Why? "Search is part of the fundamental infrastructure of the Internet. And it is currently broken," according to Search Wikia [1] (see Figures 1 and 2), the Wiki-based search engine's website. This might come as a surprise to satisfied users of Google, Yahoo!, and other search engines; however, the term "broken" does not refer to functionality, but rather to freedom, community, responsibility, and transparency.

Figure 1: With sound financial backing and a large community, Wikiasari plans to create a better search engine.

Figure 2: The Search Wikia project is looking to break up the market and make it more democratic [2].

The search engine market is highly competitive; securing a good position on the market means both revenue from advertising and direct control over the web, as German automobile manufacturer BMW learned last year when Google simply struck BMW off its index list because BMW had used doorway pages in violation of the search engine's terms of use. (Doorway pages use a technique that manipulates search engines by automatically filling hundreds or thousands of pages with popular search keys visible only to search engine bots, whereas human visitors are taken straight to the real main page.) However, the criteria that Google and other companies apply are rarely so easy to understand, as in the case of doorway pages. For example, some Google critics suspect that their pages have been pushed down Google ranking lists for reasons of content, and because only Google knows its own ranking criteria, there is no proving or disproving these allegations. Other factors, such as Google's collection of personal data and cooperation with censorship authorities in China, have led to increased worries concerning freedom of expression on the Internet while it is controlled by a few companies like Google and Yahoo!.

New Impetus for Old Idea

Thus, a free search engine can be seen as fighting for democracy, while also providing superior search results as a side effect. The project, which has provisionally being dubbed Wikiasari, is still at the brainstorming stage, although its predecessors are almost five years old. A web directory called 3apes was launched in September 2002, which more closely resembled a telephone directory than a modern search engine, and a freely editable Wiki page gave users the ability to enter what they considered to be interesting Internet addresses and categorize them by keyword. In 2004, Wales suggested setting up an improved search engine and giving users the ability to modify and improve its results. The new project was dubbed Wikia, but was soon ditched because of lack of active participation until Wales finally launched a new initiative with the same goals at the end of 2006.

Meanwhile, another hosting provider who offers web space for various Wikis took the name Wikia; in fact, the homepage of the free search engine is actually hosted on Wikia, which is why the planned free search engine is being called Wikiasari, the code name for the software behind the old Wikia project.

Wikiasari does not aim to return more results than its major competitors, but rather superior search results. Just like the battle against email spam, the main benefit of a search engine is not finding the information you are looking for, but in categorizing this information by relevance - something that Google does not do satisfactorily, according to Wales.

Wikiasari's recipe for success lies in the anticipated mass participation of volunteers who determine whether a match is interesting. Wikiasari users will be able to add to missing search results or demote top-ranking results that have not earned their ranks. The question as to whether human helpers can compete with Google's data centers is not one that can be answered until Wikiasari has motivated a sufficient number of volunteers to join the project. In contrast to Wikipedia, the project relies on making a profit through advertising, and volunteer work for a company that is profit oriented might not be everybody's idea of a good thing.

Resolving this contradiction is what Wales regards as his greatest challenge with the new project. He would like to set up a Wikiasari community that is not misused for beta testing and has little involvement in the final product, as with many other enterprises. Instead, members should become full-fledged co-developers of Wikiasari and cooperatively discuss and develop project strategies.

Nothing New?

The lack of transparency and the lack of co-determination for the user is a weak point of the major search engines. However, it is hard to see exactly how Wales' ideas of a community-based search engine will provide the quality he promises. Google also takes individual evaluations of the relevance of web pages into account, if only indirectly, by taking the number of links to a homepage into consideration. Social bookmarking pages such as digg.com follow a similar principle by letting their users collect, categorize, and browse lists of links. This said, pages of this kind are no replacement for existing search engines and, at best, can add value.

An initial version of Wikiasari is to be launched this year. In the initial phase, the initiators hope to find a large number of users who are interested in a free and open alternative to Google and Yahoo! and who are prepared to build and continually improve search engine content. This phase will probably decide whether Wikiasari ever reaches a critical mass and skyrockets like Wikipedia, which continues to attract new users, who in turn improve quality.

It doesn't look as if Wikiasari will need to face financial problems. Besides the capital that Wales is putting into the ambitious project, the new search engine will be receiving US$ 4 million from investors. This financial backing is reason enough to take Wikiasari seriously and to watch how it develops - even if it will be taking on enterprises that play in a completely different league from Wikipedia when it steps up to do battle with Google and Yahoo!.

Debian Project Leader

Every couple of years, Debian looks for a new Debian Project Leader (DPL). Voting ended on April 8, and the new leader, Sam Hocevar, is scheduled to start on April 17 [3].

The Candidates

In 2007, eight candidates ran, which beat last year's record when seven developers ran for election (see our online archives for the July 2006 issue of Linux Magazine [4]). The candidates included:

• Wouter Verhelst - Besides being a developer, Verhelst represented Debian at this year's annual FOSDEM.

• Sam Hocevar - Hocevar has been with the project since 2000.

Figure 3: Sam Hocevar, the newly elected Debian Project Leader, has been with the project since 2000.

• Steve McIntyre - McIntyre is an active member of the CD team, among other things.

• Raphael Hertzog - Hertzog ran for election back in 2002 and manages the Alioth developer service.

• Aigars Mahinovs - Mahinovs is more active in the social area of the project.

• Gustavo Franco - Franco founded the Debian Python Module team.

• Simon Richter - Richter is currently adding a degree in electrical engineering to his computer science degree.

• Anthony Towns - Defending champion Towns is also running for election.

Sven Luther's announcement that he would also be standing for DPL created quite a stir. Last year, several attempts were made to remove the controversial Luther from the project, although this did not happen [3]. Luther was missing from the final list of candidates.

Platform

In their election campaigns, the candidates promised a variety of innovations. Verhelst criticized organizational failures and promised to do everything in his power as DPL to resolve these problems. Mahinovs focused on the Debian release process - instead of regular releases, he would like to keep a kind of core distribution in a functional state to help avoid issues with obsolete software. Franco focused on Debian on the Desktop.

Hocevar adopted a slogan of "Debian sexy again." He said that creating larger working groups for the kernel areas and investing more in public relations would increase the distribution's attractiveness. Hertzog favors a DPL board consisting of developers who he considers mature enough for the position of project leader, a list that includes Verhelst, Hocevar, and McIntyre.

Richter is committed to focusing on what he thinks is a divided project on common goals and was very tacit apart from this.

McIntyre, who campaigned last year and actually ran the project as the "Second Person in Charge" along with Towns, pointed to his experience, with more transparency and a reworked maintainer process as some of his demands. In the latter case, he said more teamwork is required.

Towns' election platform started with an overview of his work in the current period of office. He admitted that attempts to convince Debian developers from various fields the merit of his ideas have shown little effect. Reason enough, said Towns, to work harder on solving problems in his second period of office.

Spoiled for Choice

In various Debian mailing lists, the role of the DPL in general was the major topic besides the question of who the future DPL would be.

Although the people who have taken part in the debate agree that the project leaders have introduced a number of technical innovations in recent years, the project's image has not necessarily benefited from this.

In particular, delays that surrounded the release of Etch have led to people calling for changes to the infrastructure Hocevar will probably be swamped with requests for changes.