Last year we focused on establishing the continuum of directories, natural language interfaces, search engines, and electronic communities. While we have some new things to add in these categories this year, the focus here is on specialty engines - databases that focus in some way on a particular topic to enhance your search experience.

The trends that I've noticed for this year include the increasing commercialization of the Web - and our attention is directed to the "Paid for location" tools. We see the increasing specialization of tools - folks really mining and displaying deep subject interests - content - that they've developed. And, we notice the different sources of information for building collections - sometimes this is the use of participant contributors for sites in a category, sometimes this is contributions of bookmarks.

We've been accustomed to information for information's sake. Now, information is being used to attract you (and your students) to other resources - in particular, things the purveyor thinks that they should buy. The problem here is the motivation for collection development: a commercial firm will do the least amount of work to convince you that this is a reliable information resource, without having the "library-type" orientation of a comprehensive collection, or a reliable collection from a particular perspective.

Increasingly, your responsibility is to teach your students how to evaluate these resources, and how to factor in commercial interests and areas of specialization. How do you evaluate the motivation behind a site, in this increasingly commercial environment? And if you're a site developer, how do you document your authority to present this information? One of the things that is so important is establishing this authority. If your prospective site doesn't have a section called "About this site," that documents who the authors are and why they can speak on this topic, run!!

In this presentation, we're particularly interested in resources for parents, children, teachers, and most particularly, the information professionals that serve them. Thus we include the information sciences resources that we should know about as professionals, as well as selected resources for your clients.

www.notess.com: Greg Notess writes for Online and Database about searching on the Web. The site includes features of major web engines and search tools, news of the engines and their capabilities, reviews of meta-engines, and a bibliography of readings.

SearchEngineWatch is also excellent for features and capabilities, and news. It's particularly targeted at web site developers interested in having their pages positioned better in the engines, and explains both policy and technical issues regarding how the various engines place a page in the retrieval display. Is the "membership fee" worth it? Yes, it's very informative and useful. His Power Searching Tutorial is very useful.

Magazine articles

Begun in February 2000 as a small dot-com in Pennsylvania, MagPortal is a directory for magazine articles on the Web. It appears to emphasize currency, and "retires" older articles to an archive (which is still accessible). It's new enough to not have a very strong database, but it's well worth keeping an eye on. The Science and Technology section includes articles from Scientific American and The Scientist.

Share your bookmarks

Quiver is a general directory of websites that began its life as Nectaris, and was renamed as Quiver in January of 2000. It builds its index by users contributing their bookmarks to the general collection; the company' s technology then organizes these and ranks them by popularity to form the database. Like Open Directory, they're using other people to build the database. If you contribute, it does collect personal information such as age and gender, and it does track your own use of your own bookmarks to tailor advertising. It's very new - nothing there on Saturn of any kind - but the astronomy links look on target. It has a personal toolbar called Qbar, but it's just for Windows.

Backflip is doing the same kind of thing: providing a personal information space by accepting and organizing your bookmarks. The service organizes them in a Yahoo-style directory. There is nothing to see if you don't sign up, however. They want to send you messages!

Top tens

Find takes a little different approach, and offers the "top ten" sites in categories such as freebies, destinations, shopping, entertainment and "big shots." The editors analyze search keywords (presumably entered at their site) and use the results to determine the "top ten" in a category. Search results are ranked by the amount paid for by the retrieved site. The highest rate paid for "insurance" was $4.00.

This is indeed a directory - more "bid for locations" are discussed below.

Directories of directories

4Anything is a directory of directories, according to its own advertising, and is heavily ad-supported. It also offers news from Reuters and features. It does offer a 4Children area, primarily for "fun" things with a few educational links.

Some standards

The Open Directory Project, owned by Netscape, represents an example of of community collaborative efforts at organizing information. It's uneven, young, enthusiastic, and international. It has also positioned itself to partner with other tools, such as Google.

Snap!, a partner of NBC, shows the differences between the scholarly and commercial approaches, in that the dot-com sites seem to get preference in displays.

Essential Links is an interesting "portal" of significant sites.

For Kids, teachers and parents

Great Sites!, from the American Library Association and volunteers contains resources for kids and parents: a collection of 700+ web resources, on arts and entertainment, history, literature, science and technology (and a little fun).

Internet Kids and Family Yellow Pages, from Net Guru Jean Armour Polly, provides a sampling from her book for kids and parents -- 100 sites in ten categories such as science, math, games, homework help, pen pals, reading and the like. Annotations are extensive and enthusiastic.

Kid Info is a "homework helper" - another term for "directory" in this context - developed by Linda Guterba, a schoolteacher in Ohio. There are sections for kids, parents, teachers, and young children. Annotations are brief; sometimes the audience (e.g. K-6) is indicated.

The Awesome Library for Teachers, Students, and Parents covers 14,000 resources. It offers different interfaces for kids, teachers, teens, and parents. The kids (and teens) interface has assembled 70 different "ask an expert" sites - from bugs to plumbers to magicians - and regular tutors.

Enchanted Learning is a purveyor of fine CD-ROM products, but also offers a collection of content resources. They offer a directory of kid-related resources, and some extensive content on dinosaurs, sharks and the like.

The Natural History Museums and Collections is an extensive collection of resources - sadly, many of the links are broken - but as an international collection, it's worth taking a look.

Smart Zone from Edview is a catalog of "teacher-approved" resources, and they offer "channel lock," a piece of software that works with a browser to prevent the browser from looking at any resources outside of the Zone. Searches can be targeted for elementary, middle, or high school. The Saturn search turned up 11 astronomy sites for the elementary set, 2 games, and 1 car.

Study Web is another directory from American Computer Resources. Launched in October of 1997, it covers 88,000 "research quality" sites, selected by its "faculty" - which appear to be volunteer contributors. It appears that a few people contribute to a given area, functioning like editors for an area. Entries have summaries, indication of visual content, grade level appropriateness. Coverage is not comparable in breadth to Yahoo: Nanotechnology has 32 entries to Yahoo's 75: but you can figure out what you want to look at more easily here.

The State of Tennessee offers its Kid's Pages, which include quizzes on state government, advice on how to be a healthy and safe kid, and a money game.

The KIDS (Kids Identifying and Discovering Sites) Project is a part of the Scout Report. In the KIDS project, K-12 students review sites that are of interest to them, for other K-12 sites. Each bi-weekly newsletter focuses on a specific topic.

Kids Web, from the Northeast Parallel Architectures Center at Syracuse University, is a directory of resources for K-12. Annotations are brief, but the subjects include the sciences, arts, social studies. Sports, Games, and reference materials are in the "miscellaneous" section.

KidsClick! a yahoo-type directory of resources by librarians, with annotations, reading level, and presence of illustrations. Of particular interest to kids will be the "Weird and Mysterious" category.

AskERIC, the interface to the DOE database, should be familiar.

A new engine, FAST at AllTheWeb, claims to search 300 million pages now, and will keep growing to cover the entire Web of over a billion documents. The kudzu search yielded 20,000 documents in less than a second.

Launched in March of 1999, TopClick is a search engine powered by Google, but is designed to protect user privacy. It keeps no records of your searches, and uses no user-tracking capabilities. A search of "kudzu" proved to be exactly the same, but without access to the cached copy of the site or GoogleScout. The site also contains no banner advertising.

Meaning-based engines

Some of the newer engines are using lexicons and thesaurii to improve retrieval effectiveness. The Saturn search is an example of the classic problem - traditional engines retrieve a mix of meanings. At Oingo the Satur n search presents mixed-meaning categories, and mixed-meaning websites, but also the opportunity via a menu to narrow the search to the planet, a town, the car, and the game.

In February of 2000, Simpli.com launched SimpliFind, an engine that can be supported by meanings that you create. Using "interactive query disambiguation," multiple meanings can be managed. The search box allows y ou to enter a term (Saturn) and before the search is conducted, clicking on a menu causes the dictionary to be searched. For Saturn, the meanings of car, game, and planet, were presented - along with the opportunity to enter a new meaning. It's a meta engine - results were found by Google, DirectHit, Excite, on the first page of results. The advanced search just allows the opportunity to add more terms. I added the term "planetary" to the meanings for "rings" - and the results were very targeted.

The third of these is Ejemoni, which has no working model yet, but is something to keep an eye on.

Bid-for-location

We are increasingly seeing "portals" - combinations of search engines, directories, and general information - using "bid for location" services to rank retrieved hits. A few years ago, OpenText's proposal to do just this was received with horror in the L IS community. It seems to be accepted practice now. Kanoodle is an example; its highest bid for "insurance" is thirty cents. It offers a filter, but the unfiltered results are quite racy. Like GoTo, Kanoodle depends on Inktomi for its unpaid listings.

GoTo is simply selling site positions, first shown to the highest bidder, as Alta Vista also is now doing.

Rocketlinks is another, offering news, entertainment, "top clicks," "cool clicks," all advertiser-supported. The top "bid" for "insurance" was seventy-one cents.

FindWhat is yet another: its top insurance bid is ninety-one cents. It relies on inktomi for non-paid results.

HitsGalore is another portal, with categories, news and entertainment information. Its top "insurance" bid is $3.51. It claims to offer porn-free searching - a search for "sex" proved to be mostly advertising for Viagra, and a motor components company.

SimpleSearch just sells the first five positions. Other "bid for location" services have no limit on the number of positions they'll sell. Unpaid hits are provided by DirectHit. It is significantly less cluttered than the others; its Learning Center offers instruction on internet and computer topics. Its highest "insurance" paid entry is nine cents.

Find takes a little different approach, and offers the "top ten" sites in categories such as freebies, destinations, shopping, entertainment and "big shots." Search results are ranked by the amount paid for by the retrieved site. The highest rate paid for "insurance" was $4.00.

Paying you to search

IWon is another portal, offering news headlines from Reuters, consumer-oriented categories, and entertainment such as horoscopes. Each link is assigned a point-value, and you collect points as you navigate the portal. If you collect 300 points in a week, you're entered in a drawing to win a prize. This is definitely a manipulation of your Web experience. Its database is from Inktomi.

Revisions to the majors...

AltaVista has released its new engine (3.0) as Raging. REVIEW

Searching is settling in as an industry, and this year we're seeing major revisions to the major engines. These include partnerships with other tools, and inclusion of directory-type information to supplement the computer-generated tools.

Alta Vista has implemented "search centers" - specialized areas for narrow topics such as MP3 files and images. Image search also allows for search of "partner" databases such as Corbis. If you're interested in how engines work, the company has provided detailed information on its ranking practices.

Google has added a directory. The sites included are listed in the order of Google's analysis of their popularity. The directory base comes from the Open Directory, the link analysis which affects the order in which items are displayed comes from Google's popularlity rankings. This means that you can go to a category and turn your search into one narrowed to the category: searching the category home:houses for windows will omit the OS. You can read more about this at Salon Magazine. Google has also added a language search: you can specify the language you wish with a pull-down menu. I searched for "bibliometrie" and specified French, and it retrieved just sites in French.

The site has also added a specialty search engine for Macintosh users. The logo for Google sports the Macintosh colors.

You can now "personalize" your own newspaper at AskJeeves.

In a variation on the theme, ATT WorldNet is also using the Open Directory Database, but using DirectHit technology to rank the listings.

SearchEnginesGalore has morphed into MetaIQ earlier this year, and now functions as a meta tool, which specialized subject engines, links to ask-an-expert services, free e-mail, an MP3 engine, news of the day, and more. It appears to read in the Open Directory Project information, but without credit to the editors. Its search results do tell you where the individual items came from.

New interfaces are presented at SavvySearch (with a new mission) and Proteus.

Ranking mechanisms are changing for some engines, particularly those with priced rankings: See Hotbot for its relationship with DirectHit: a compilation of sites that others doing the same search found useful, or at least visited. Google is tracking other sites that link to a given page as a measure of utility.

GlobalBrain attempts to go beyond human directories and tap into the preferences of users themselves by tracking popular sites and integrating the results into the retrieval algorithms.

Engines for kids...

Kids have their own meta tool at KidsSearch, developed (apparently) by the Ramapo Catskills Library System (they seem to be a bit shy about authorship). It combines search boxes for AOL Netfind, KidsClick, the Awesome Library, Searchopolis and others. It includes a selection for parents and teachers, as well as a "just for fun" section. The site could be a bit neater.

AOL Netfind | Kids Only offers a search engine and set of fun things to do (eg solving Nancy Drew mysteries and safe chats). Saturn yields mostly the planet and related astronomy links - but a few cars, and the wrestler shows up!

Super Snooper includes only sites which "do not contain information of a pornographic, violent or hatred-related nature." Cool Sites Network believes in the ability for people to access any types of sites they wish: but, also believes that other sites provide this access. They have cataloged (as of an unspecified date) 1.5 million sites, primarily from the English-speaking world. The saturn search turned up 100 entries, not a single one from NASA> It appears to be searching the URL's first - about 90% had saturn there. Almost all were Saturn dealerships, with a few game sites.

ISleuth offers "windows" into the major engines as a search aid, as well as directory capabilities.

Multi-media

Multi-media engines include Alta Vista's Photo Finder, and Scour.net. The first is very concerned about copyright issues, the second doesn't seem to be.

Research projects

Research projects are focusing in part on clustering. The IBM Clever project is trying to develop an "Automated Resource Compilation," understanding both bibliographies and sites of significantcontent. GROUPER and HuskySearch work together to cluster results of a search by examining words in common in retrieved pages. Informedia Mediator is from a Swiss Bank, and offers interesting options in customization.

For children, Ask Jeeves for Kids includes resources from other engines and directories, but filters them with SurfWatch. The "Where is Saturn" search yielded planetary information from their database, but cars from the external links!

Most of the major engines are accepting natural language, queries, but your results are web sites. At ESPN StatSearch the idea is to enter a question, and get back an answer. Reviews have it tripping over "best" - retrieving sports players, rather than identifying the "best" player at something. Its sample questions aren't very natural language, despite the suggestion directly under the search window. At the moment, it is limited to Men's basket ball.

Conversational interfaces ala Star Trek are on the horizon, and making strides from ELIZA. The Brain is not quite the Enterprise computer, but it's an interesting step forward.

But there's a new one I want to bring to your attention, that works a little differently. It is still "people to people," and still "expert to questionner." But at ExpertCity, you can type a query about a problem with your PC into the search dialog box, and several "experts" will in real-time bid to answer your question. You can choose your expert based on their fee (usually in the $2-$20 range, according to their documentation) and their skill level. If you've signed up for their service, you've already downloaded and installed the plugin that enables the "expert" to "see" your screen. The software has the capability to let the "expert" control your keyboard and mouse, if you allow that. It's only for Wintel machines at this time.

While enature is technically not a general search engine, and is limited to the US, its treatment of its database of nearly 5,000 creatures from bugs to birds to seashells is spectacular.

Studying fish? FishBase may help. While badly documented (and some digging suggests that these folks know what they're talking about), this site presents a detailed approach to our fishy friends. The photo gallery is great preparation to a visit the aquarium. Searches by scientific name as well as popular name are enabled, as are quizzes and the opportunity to become a collaborator.

If Magic is your interest, visit MagicTricks Library - an extensive database of things about magic and magicians. It includes program information, celebrities, women in magic, magic on film, trivia, and publications. Yes, this is a magic shop - but it's clearly put some time into gathering all of this information together.

Students doing research on public figures in sports, politics, entertainment, the news, crime, religion and science may find useful information at PublicFigure, a database of in-depth profiles of 5,000 individuals and briefer information about 17,000 more that appears to have launched late 1999. Profiles include brief bios, lists of works, photos and news, and a profile. It isn't just people living today: Carl Sagan had a brief profile, however Dylan Thomas did not. Bob Dylan did, Mama Cass did not. Neither did George Washington, but Thomas Jefferson did. Margaret Thatcher did, as did Vladimir Putin, Gandhi. The selection criteria are not given, but there's an extensive discussion of the definition of a "public figure," relying in part on Supreme Court and other legal decisions. Beyond these profiles you can purchase books, videos, and memorabilia of the persons included.

In August of 1999, Hungry Minds launched to bring together a variety of ways to learn just about anything: subject experts, experts for hire, online courses, degree programs, books and other media. Topcs are presented yahoo-style; a search engine searches across all categories. If you're an expert in an area, you can join them! A search for "tomatoes" turned up some interesting recipes, but no "how to plant" guide. It did turn up an online course on how to harvest and store them, with a description of the course, an instructor bio, and customer reviews of the course. A search for "basketball" turned up a wider variety of resources, including a $158 course from Oklahoma, and some free guides and tips. This site lead me to eHow, a collection of "how to do it" instructions - "make a paper airplane" is excellent! The site includes contributed user tips and an opportunity to buy things you may need to complete the task. If you want to be reminded to do the task on a particular date, it will send you an email!

Interested in finding the Website of a museum? MuseumNetwork can help. This is a directory of museums around the world, including ones for children. The museum database can be searched directly for one that you know of, or, you can see lists of them categorized by broad categories such as art and design, history, science and nature, and sports. Science and nature is broken into subcategories such as aquariums, bird sanctuaries, earth sciences, natural history and the like. I'm not fond of the pop-up windows, however they will "hang around" when you're looking at sites.

Interested in Theater? The Internet Theater Index is a directory of theater resources - performances, performing companies, practices (acting, playwriting, singing), and education (including drama schools). It launched in February and has added 83 links since then - looks like a slow grower.

Studying the tides? Then HarborTides is the place to find tide information, sunrise and sunsets, and high and low tide heights. Search by location or zip code.

Looking for information about a contemporary artist? About a painting? ArtStar Explorer may be able to help. You can find artists by name, but also medium, material, period, and subject of their work.

While it's important to teach kids about intellectual property and when they can use the images they find on the Web on their own sites, it's also helpful to steer them to resources they can use without question. At GoGraph they'll find animated flowers, dragons, stars and planets, and a host of other images for personal use, organized in a Yahoo-style directory. The associated search engine also appears to be concept-based, not just based on file names.

In addition to scour.net and the AltaVista photo finder, you now can use ditto.com to search for images. Humans have evaluated the images, and indexed them. These are images found on the Web - they may be a part of a magazine article, or a report, or in a collection - you must contact the original owners for use. Navigate using categories such as Animals, Nature, Holiday, Sports, Space and others. Or, you can search. The search for Saturn turned up the planet, the car, a couple of books, and the wrestler. Alongside of your image results is "ditto shopping," wherein an equivalent search has been conducted on things available for purchase. If you can't find what you need, the Ditto Detective will look for it for you.

Since May of 1999, FasTV has offered categorized (news, sports, entertainment, business, lifestyle, politics) and searchable video clips from sources such as CNN and Jefferson-Pilot Sports. It has had exclusive Web coverage of major sports events such as the PGA, and SEC sports. When I worked with the site in April, every time I returned to the News page, a different story was presented. Be careful of the CNN designation "live" - these are clips, not live feeds from CNN. Video is streamed through the Real Player.

The national election is upon us, and specialized tools (beyong the candidate's own efforts) have emerged. Northern Light offers links to news, political sites, and a search of their database. Search across the candidates sites for your issue at ElectionSearch2000. A similar service is offered at Issues2000, which organizes the candidates and their issues. On the 25th of April, Excite launched Excite Elections, a guide to elections around the country, with lots of interactivity in the form of polls, message boards, clubs, and news. You can compare candidates here in "My Elections."

Danny Sullivan offers an analysis of the relative "finability" of major candidates on 18 major engines, and offers tips on making your site more retrievable, at Can You Find Your Candidate?.

SearchGov is a directory and search tool for US federal government information. Contrast this with FedWorld.

See also Sports, above in Natural Language Interfaces.

You've probably read about attempts to access the Internet on your cell phone, pager, or other small device. These devices can't swallow an entire Web page - they need specially-created files. WAP (Wireless Application Protocol) search engines are designed for display on small devices such as a cell phone, Palm Pilot, or Cassiopoeia. These engines send you to WAP sites, which are intended for viewing on such devices - rather than large Web pages which are designed for monitors. See Fast for a search engine, and WAPLY for a directory. The latter includes Web sites about WAP technologies.

Indexing speech? SpeechBot is experimental index of radio shows from Compaq.

A new (2000) search guide is at Lookoff, with a good set of "do's" and "don't's" and an opinion about the best tools to use.