October 2006 Archive

Open Content Archive: Books Available

Since the announcement of the Open Content Alliance (OCA) back in Oct. 2005, I have been waiting to see the results of the project and view some of the scanned books. While the major search engine partners like Yahoo! and Microsoft have not yet launched any product, I noticed that the Internet Archive does have quite a few books already available in their Text Archive (over 30,000). Look at a record to see the rich meta data included (not just title, author, date, and publisher, but also language, contributor,call number , scan date, and image count). The book is available in PDF, txt, DjVu, ftp, and Flip Book formats. The Flip Book is a nice interface showcased at OpenLibrary which mimics the flipping of pages in printed books. So look for OCA books at the Internet Archive's Text Archive. It will be interesting to see if all of these books are available at all the partners sites, not to mention comparing coverage to Google Books and Amazon.

Yahoo! "Open" Shortcuts: Quick Navigation with !

Yahoo! has announced more "Open Shortcuts." The Yahoo! shortcut commands are preceded with an exclamation point and entered into a Yahoo! search box. For example, enter !wiki asbestos at Yahoo! and go straight to the Wikipedia entry for asbestos. Enter !list to see a list of the preconfigured shortcuts. The open shortcuts are ones that can be user configured and tied to a specific Yahoo! login. See the blog post for more instructions and links to the various ways to create these.

Subset Searching at Google

Used Google Co-op much? No? Other than developers and experimenters, I have heard little use of Google Co-op, especially for searching. That has all changed with today's launch of Google Custom Search Engine, an application of Google Co-op. The Custom Search Engine was announced yesterday.Like other search engines that search a subset of a larger database, Google's Custom Search Engine lets users specify specific sites to included or excluded, and some can be prioritized over others. Creators can also specify that certain words should be added to the query. Other subset (or vertical or custom) search engines include Yahoo! Search Builder, Rollyo (which uses the Yahoo! database), and Eurekster's Swickis

Weekend Search Changes

This past weekend, I've been giving several workshops in Monterey at the Internet Librarian conference. It is always fun to give a workshop on Web searching on the weekend when the search engines tend to roll out new features or give otherwise unusual results. I had two such situations this weekend, both of which are back to normal today. First, I tried to demonstrate searching the Yahoo! directory from the main Yahoo! page by just clicking on the "Directory" tab above the search box. On both Saturday and Sunday, when searching "monterey" that way, Yahoo! said there were zero directory results, even though if I went straight to dir.yahoo.com and search "monterey," I got plenty of results. Meanwhile, over at Google, I was demonstrating what happens when you click on the "more >>" link which used to bring up links to "Books," "Froogle," "Groups," and "even more," but no link to Google Scholar. On Sunday, "Scholar" had been added to that last of four. I was hoping that the Scholar addition would be permanent and that Yahoo! would fix the directory search. As of today when I tested both situations, the Yahoo! directory searching is fixed, but Scholar no longer appears under "more >>." Oh well.

Cornell in for Live Book Search

Both Live and Cornell have announced that Cornell University is joining the Live Book Search project. As with such announcements from Google Book search, that means it will be awhile (perhaps a year or more) before any books from that library are available online. For that matter, Live Book Search is not yet available, although it is supposed to go live (no pun intended) later this year.

LinkFromDomain: New Advanced Search Operator at Live

Live has a new advanced operator. This is an all to rare occurence, so I'm happy to see it available. The Live Search Weblog post describes the new LinkFromDomain: field search operator. Basically, for a specified domain, this will find what external links the pages on the site point to. The command can be all lower case as in linkfromdomain:notess.com or with initial caps LinkFromDomain:notess.com. It can also be combined with other field searches. linkfromdomain:loc.gov site:hr finds Croatian links on the Library of Congress Web site's pages. Danny takes an in-depth look at it as well.

Interesting Ratings of Consumer Satisfaction with Search

Guess which search engine has the highest satisfaction scores in a J.D. Power and Associates Report? Their 2006 "Residential Online Service Customer Satisfaction Study" surveyed 10,787 U.S. residential Internet users and found several surprising results. "The study finds that nearly 75 percent of Internet subscribers use multiple search engines. Forty-four percent of those customers report using multiple search engines because each one has better sources depending on customer needs." While the study echoes several others in its finding that Google is the primary search engine for a majority (but only 51%), the customer satisfaction ratings give the top spot not to Google, or Yahoo!, or Ask, but to Dogpile. On a 1,000 point scale, here are the search engines ranked and their point values:

1. Dogpile 804
2. Ask.com 793
3. Google 784
4. Yahoo! Search 756
5. AOL Search 742
6. MSN Search 736

Ask Mobile

Ask announces the launch Ask Mobile, a cell phones and other mobile device version of Ask.com Web search. The posting highlights numbered shortcuts for quick access to Ask tools, a search history of recent searches, and the formatting of Web pages to better fit on the mobile screen.

New Exalead Interface Launches

Exalead has launched its new interface that has been in beta and preview mode for the past month or so. They also report indexing over 8 billion pages (they had initially stated they would meet that goal in the summer, so they are not too far behind schedule). Exalead Review has been updated.

    See more »

New AOL Search Beta

Pandia has a report on the AOL Search 7.0 beta and how it compares to Google (from which AOL gets its search engine and advertising). AOL continues to receive a healthy dose of search traffic from its own users, and it does offer different features than Google.

A Comparison of BitTorrent Search Engines and Critique

An interesting blog posting today complains about another well-referenced blog post that claimed Btjunkie is the largest BitTorrent search engine. What's BitTorrent? BitTorrent is a peer to peer file sharing technology often used to transfer large files, sometimes of copyrighted material. Since I am not a frequent BitTorrent user (having only checked it out quickly a few times), I will not give any personal preference to these search engines.What struck me the most about the post is that it has an excellent critique of search engine size studies. Basically, it is easy to get a larger number of hits by leaving many dead links in the database. That reminds me of a tale I heard once from an AlltheWeb engineer who talked about finding millions of pages of German spam in their database and then removing them. The next day he heard complaints from some German users because the number of results on the searches they were testing suddenly decreased. The German users assumed that AlltheWeb was crawling less of the German Web, even though from the engineer's perspective, they were just cleaning up the database.

FreeTechBooks.com: A Directoryof Free CS and Programming Books

FreeTechBooks.com is a nicely organized directory of freely available programming and computer science books and lecture notes. The free books listed are usually available directly on the publisher's Web site. The classifications on the left side includes books in Computer Science, Operating Systems, Programming and Scripting, and Related Fields. A search box will search title and descriptions of the books but not their full text. Based on the numbers per category, it looks like it includes about 100-200 titles. While some of the titles are available via sources like The Online Books Page, others are not included there.

MSN Search Not as Dead as I Thought

Back on September 15 I noted that MSN Search was gone, fully replaced by Live Search on all of my searches, at least. However, today I found some exceptions. I had only been using computers with Windows XP. Using IE6 on a Windows 2000 computer today, I found that I could still search at MSN Search with the old MSN Search interface when using the browser search button (assuming that the button is still set with the default of MSN search).

    See more »

Searching Computer Code

A new specialized search capability from Google Labs is the Google Code Search, in beta. Unlike regular Web searching, the Code Search allows truncation with the asterisk (*) and regular expression searching, but it is only searching across identified programming code source like JavaScript and Perl. According to the announcement, it is designed to give "programmers a single place to search publicly accessible source code." It has some specialized field searches including file:, package:, language:, and license:.

New Google Groups

The old Usenet groups and DejaNews search engine became Google Groups. Now Google is announcing the launch of yet another version of Groups. The beta has a new interface and several new features. New capabilities include the ability for group owners to create a welcome message, upload a group logo, and customize fonts and colors. The new Pages feature lets users create web pages inside a group as well. Overall, it seems to be moving Google Groups further away from its Usenet origins and more towards what Yahoo! Groups offers.

SearchMash, the New Google?

Google usually runs all kinds of experimental user interfaces, changes to ranking algorithms, and tests of every type on their own site, just letting a very small percentage of their users see the changes and then evaluating the response. Now, Google has actually launched its own experimental search engine a a completely different URL with no Google logo. SearchMash is the new site. SearchMash presents Web results with Image matches off to the right side. Results are numbered (something I always prefer) and can be sorted. Just click and drag one to change the order of the results (although I have no idea why you might want to do this). Another interface change is that at the end of the page, searchers can just click on the "more web pages" link to retrieve another ten results. Instead of being shown on a new page, they just appear below the first ten, on the same page. Overall, I find no compelling reason to use this, but it is a site to watch for future Google experiments.