|
|
|
|
| | ||||||
| |
|
|
|
|
||
|
Home Page Search this Site Pandia FAQ Search tools: Plus Web Directory Metasearch Newsfinder Shopping Search Radio Search Powersearch All-in-One People Search On Searching: Search Tutorial Books Resources Search Engine News Syntax Q-cards Free Newsletter
|
The Pandia Post Newsletter No. 10September 2001
In this issue: EDITORIALThe horrible tragedy of September 11 has marked us all, and the September issue of the Pandia Post will be a short issue. Now should be a time for reflection. An important part of any grieving process is the need for information. We look for clues and patterns in order to make sense of the senseless, and find meaning in the meaningless. The Internet played a very important role in this respect. One of my colleagues found out about the first plane by visiting an online newspaper, and soon we all searched out various news sources in order to get more information. CNN.com was unavailable parts of that day, as was some of the national online newspapers. Even the major search engines struggled under the burden. According to the BBC, traffic to search engines and portals was up by about 10 times more than normal. The Pandia statistics log program actually gave up under the pressure, but we have indications that the number of visitors to our online radio directory went up some 1000 percent. People has been looking for "online radio", "online radio news", "web radio" or just "news". The most popular news organisation searched on Lycos and the Google search engines was CNN, followed by the BBC, MSNBC, and ABC News. The fact that no search engine operates in "real time" is a problem. Spidering the Web and updating databases take time, often weeks or even months, and few of the search engines could present relevant links for "World Trade Center", "Twin Towers" etc. right after the disaster had struck. If you pay for it, a search engine like Fast will spider your site every 24 hours, Inktomi every 48 hours, but even this is too slow when events are unfolding minute by minute. Google tries to compensate for this by spidering news sites like CNN at regular intervals, but as Danny Sullivan of Search Engine Watch has pointed out, searches did not produce any relevant results as late as two ours after the first impact. On the other hand, a couple of hours later, Google had spidered the CNN site, and provided "cached" copies of the site for searchers who were unable to access CNN directly. This is certainly problematic from an intellectual property point of view, but we are quite certain CNN did not mind. AltaVista, which -- like the Pandia Newsfinder -- gets news from the Moreover news service, managed to produce relevant listings much earlier. At the same time a lot of smaller sites set up special pages with links to relevant news resources, public institutions, hospitals etc. Pandia Award Winner Tara Calishain turned her ResearchBuzz site into a virtual resource centre for disaster information. The page is still there. Some unknown symphatizers in Denmark put up a special site, where people from all over the world could express their feelings by lighting a "virtual" candle. At the moment of writing some 350,000 candles have been lit. All democracies were under attack this week. The Danish site demonstrates that we are in this together.
Per and Susanne Koch Search Engine Watch on finding disaster coverage at search engines: http://www.searchenginewatch.com/sereport/01/09-wtc.html Changes at Pandia:There have been no major changes at Pandia since the previous issue of Pandia Post. We have removed some dead links at the Pandia Powersearch all-in-one search gateway, and added a few new resources. We have also extended our news coverage in the search engine news section somewhat. Previously we used to include links to separate Pandia news stories only. Now we might also present small news snippets that does not require more detailed analysis. In this way Pandia Search World will function more like a search engine news web log. Pandia Powersearch: http://www.pandia.com/powersearch/ |
|
Click here for eBay! |
PANDIA SEARCH WORLDThe ambiguous Inktomi anti-spam policyWebmaster World remains the most up to date bulletin board for professional search engine discussion. Brett Tabke, the editor of both the Webmaster World and the sister site Search Engine World, publishes a newsletter called Search Engine World Quarterly Magazine, and both Webmaster World and the latest issue of the newsletter focus on the Inktomi anti-spam policy. Clearly Tabke and many of the debattants fear that Inktomi is discriminating against search engine optimization companies that do not take part in Inktomi's various paid inclusion programs. The debate is full of serious allegations and few definite proofs, and we are still waiting for a response from Inktomi. Still it demonstrates one thing clearly: The integrity of a search engine company like Inktomi is suffering because of the new paid inclusion services. Just to mention one point here: Inktomi asks websites to pay for a regular 48 hours spidering of their sites. In one of the programs, websites with more than 1000 pages may specify which pages are included in the index and decide when the pages become available. In that particular program the sites pay a certain sum per click through. But what happens to the sites that do not take part? According to the Inktomi policy websites may be included in the search engine index for free, but the listings will be updated with a much slower frequency. However, many of the debattants are not convinced that this is so. Jeremy Goodrich, one of the participants in the Webmaster World discussion, says that "Inktomi pulled everything that this company had, and traced everything they had from the domain registration database, and four c classes [i.e. IP web addresses]. This was how they got black mailed into becoming an Inktomi partner...Ink took away their Ink derived revenue, because they knew which urls to take out, and which c classes, etc. and then, 3 months later to remind the company just how valuable inktomi is they let them have their traffic back for about a week. Just a taste, to see what it would be like...and then they got cut off again. " Pandia has found no proof that Inktomi is systematically blacklisting search engine optimization companies in order to force them into taking part in paid programs. Maybe some of these companies have practiced spam (i.e. Unsavory techniques used to trick the search engine into giving webpages good rankings), and maybe Inktomi lets them into the paid programs as they feel they better can control spam in this way. The fact remains, however, that Inktomi is loosing credibility in the search engine optimization milieus, which ultimately may undermine its legitimacy as a reliable and "objective" search engine. It cannot afford that, in a market that is constantly killing online search services. On the other hand, it cannot afford NOT to get revenue like this. It's a terrible dilemma, indeed! Read more about the discussion and the Inktomi anti-spam policy in Pandia Search World: http://www.pandia.com/sw-2001/54-mediadna.html GoTo changes name to OvertureThe pay-per-click search engine GoTo will change its name to Overture on October 8. We find it hard to believe that anyone would let go of a snappy and well known trade mark like GoTo, but as GoTo is one of the very few profitable search companies, we suppose they have to use the money on something. Pandia has more on the new name: http://www.pandia.com/sw-2001/52-overture.html Two heirs to GoWhen Disney dropped the Infoseek search engine and the Go guide directory, many of the old directory editors felt bad about going into early retirement. Instead the now defunct Go directory has given birth to no less than two new directories: JoeAnt.com and Goguides.org. Like the Open Directory (which feeds the Pandia Plus Directory) and Zeal these are edited by unpaid volunteers. Goguides.org is organized in the traditional Yahoo!-way, but there are not that many listings yet. Websites added to the Goguides Directory should be quality sites that "add original content and value to the directory", the site says. Quality content seems to be the key. At the moment it seems JoeAnt is shaping up to become the most serious contender to the already existing directories. This is also a Yahoo!-like directory, although the site goes for a pull-down menu based navigation system instead of the traditional category lists. Every listing is followed by various icons giving information on the site (like $ for "shopping"). An Info-button brings up a page where you can search for the owner of the site and search the site's content using Google. When using the search form you may select "All Words" (Boolean AND), "Any Word" (Boolean OR) and "Exact phrase". We could not find any help page. |
|
SITESEEINGWiseNut -- the new Google?The NBCi and Go directories are dead, as is the Infoseek search engine. Excite has also a rather pale complexion these days. That's why it is so good to read about new and exciting search engines like WiseNut and Teoma. After all, Google and Fast came out of the blue, maybe these newcomers will be the future winners of the search engine race? Many attribute the success of Google to its PageRank system. The more inbound links a page has, and the higher the importance of the pages linking to it, the higher its Page Rank. Google also uses page internal factors when determining the ranking of a webpage, i.e. how many times the keyword phrase is present on the page, and -- above all -- where the keyword is situated, in the title-field, in the headlines, in the first paragraph of text etc. Nevertheless, what made Google such a huge hit, was their use of link popularity. The logic goes like this: If a site has a lot of inbound links, that is for a reason. Most likely the site contains information that other webmasters find useful. If a high quality site links to your site, that is likely to mean that your site is also of a certain quality. Any link won't do, however. You get a much larger boost in the rankings if the content of the page or site linking to you is of direct relevance to your page or site. You can determine the relevance of the link in two ways. You can see if text in or near the link is similar to the text on the page the link refers to. If this text also fits the search query entered by the searcher, you are getting closer to a good ranking. Google has also been using so-called theme based indexing. This means that it tries to analyse the whole site to see if it has a consistent theme or topic. If nearly all the pages of a site is dedicated to -- let's say -- barock art, the chances are that this is an expert site. If a lot of sites devoted to barock art points to this site, you can be quite certain that it is a competence hub. And if your site on Bernini gets a link from this expert site, you are in luck, because that means that the experts have found your site valuable. Google likes that. Today, all search engines used techniques like this, more or less efficiently. The WiseNut ranking system considers three levels of context when evaluating relevancy: 1) all the words on a Web page, 2) the text of referring links and the words around them, and 3) the significance and content of the pages containing the referring links. WiseNut claims it has a fundamentally different kind of relevancy ranking system compared to Google, as it uses the text of hyperlinks, the words surrounding them, and the structure of the Web to evaluate and rank pages. It is a context sensitive link analysis system that not only measures the relative importance of a particular page, but also determines the relative relevancy of that page for a given query. This new technology is virtually impervious to most forms of spamming, WiseNut says. It should be noted that Google associate the text of a link with the page the link points to, so why WiseNut's ranking system is so fundamentally different is a bit hard to understand. That should not concern us too much, however. The most important thing here is that they have developed a link popularity oriented search engine that actually works. Whether it is as good as Google, is a matter of debate (and -- maybe -- taste), but it does deliver highly relevant results. Moreover, it has succeeded in building one of the largest search engine databases in the world. The search engine index contains some 800 million pages. Only Google, with its 1 billion pages, is larger (at least until Fast launches its 1.8 billion webpage index this fall). What we like best about this search engine, is the WiseGuide categories, which are a bit similar to the Excite Zoom function. At the top of the page you get a list of alternative keyword phrases that may guide you to even more targeted search results. These semantically related queries are actually generated on the fly, and can be very useful when tracking down hard to find information. Like most new search engines, WiseNut, is a disappointment when it comes to advanced features. There is a Preference page, that lets you set preferences for the number of displayed results per page, result clustering, languages, and a content filter. There is, however, no regular help page and no support for true Boolean searching. By using the separate "WiseSearch" page, you may use a search form to designate keywords that must be present (+), must not be present (-) and additional exact phrases (" "). WiseNut also accepts search engine math, i.e. the use of the signs +, - and " ". (For more information on search engine math, see http://www.pandia.com/goalgetter/10.html.) WiseNut http://www.wisenut.com/ Teoma adopted by Ask JeevesAsk Jeeves, a search engine company known for its ability to interpret natural language queries, has now acquired the privately held Teoma Technologies, Inc. Ask Jeeves will integrate Teoma's search technology across its Web properties, including Ask.com. It will also integrate the Teoma technology in its corporate search and syndication products. Today, companies like About.com, iWon, MSN and Lycos use Ask Jeeves technology. Like WiseNut, Teoma is a site that tries to follow in the steps of Google, presenting a popularity based search engine within the framework of a clean and simple lay out. It has a much smaller database than WiseNut, though. It contains only some 100 million pages. Teoma was started by scientists at Rutgers University, and the search engine has been hailed in recent months as "the next big thing in search engines". The name derives from the Gaelic word for "expert," and the search engine determines results by ranking a site based on its subject specific popularity (the number of Web pages about the subject that reference this page) as well as its general popularity (the number of all the Web pages that reference this page). Teoma also presents so-called "communities" of expert sites, i.e. relevant knowledge hubs that may guide you in your search. This is a very useful feature indeed, maybe even more so than WiseNut's WiseGuide categories. The fact is that most professional searchers are on the lookout for expert sites that may guide them around in the webscape. Teoma tries to give them that. Still, it should be noted that many of these are not so much expert sites as they are directories or lists of relevant sites. Here's a tip: Webmasters can use the Teoma community sites to find potential linking partners. Teoma also has a directory like autoclassification of web pages. Near the top of results pages you will find a section called "Web Pages Grouped By Topic." Here all the results have been grouped into broad categories, just like in Yahoo! or the Pandia Plus Directory. According to Search Engine Watch Teoma analyses at the results set, then seeks out "clusters" or "communities" of 300 or more pages that link to each other. When these clusters emerge, the link text is analyzed to find the most common words, which are then used to describe the category. Google has considered a similar feature, but chose to include their own variant of the Open Directory instead. Like WiseNut, Teoma has only rudimentary functions for advanced searching. Yes, you may use search engine math (+, -, and "phrases"), but that's about it. It defaults to an AND search (i.e. all words must be present on result pages) and you cannot connect words with a Boolean OR. There is no search form for advanced searching. Still, Teoma is in its beta testing phase, and who knows what will happen once it has been integrated into the Ask Jeeves website.
Teoma: http://www.teoma.com/ |
|
Pandia -your guide to search engines |
SEARCH TIPSAllthweb.com advanced searchFast's AllTheWeb.com search engine has been given a successful face lift. Although we know that Fast has been planning the introduction of "real" Boolean searching, this is yet not there. You still have use the pull-down menues on the page for advanced searching to bulid more complex queries. In so far as Fast has implemented more complex search operators, these are -- unfortunately -- nonstandard. You may for instance search for any pages containing the word "cat" and/or the word "dog", not by using the OR-operator -- which normally is the case -- but by using brackets. Hence the query -- pets AND (dogs OR cats) -- ends up as -- pets (dogs cats). Our friends in Trondheim, Norway, should be able to do something about this. There is also a new list of field operators, i.e. operators that restrict searchers to particular sites or parts of webpages. These are: url.tld:domainname -- will find pages within the specified domain. url.tld:no -- will find pages from Norway. url.tld:com -- will find pages from commercial sites. url.host:name -- will find pages on a specific site, hence url.host:www.pandia.com will find pages on the pandia.com site. link.all:URLtext -- will find pages with a link to a page with the specified URL text., hence link.all:www.pandia.com -- will find all pages linking to www.pandia.com (which is very useful for webmasters trying to ascertain a site's link popularity!). url.domain:text -- will find pages with the specified word or phrase anywhere in the domain name (like pandia.com). normal.title:text -- will find pages that contain the specified word or phrase in the page title (which appears in the window bar of most browsers), hence title:cats -- will find pages with the word "cats" in the title. url.all:text -- will find pages with the specific word or phrase in the URL, hence url:dogs will find all pages on all servers that have the word dogs anywhere in the full URL (e.g. http://www.dogs.com/ would be in the results). normal.titlehead:text -- will find pages with the specified word or phrase in the title or in its head content (including metatags, presumably). link.extension:text -- will find pages that contain links to files with this or that file extension, hence link.extension:jpeg will find pages that contain .jpeg images. Explorer AutoCompleteSeveral of our readers have asked us how to cancel the "autofill" feature of the Pandia search forms. This feature tries to guess what you are searching for by completing your search query on the basis of previous searches. This is actually not a feature particular to Pandia. It is the Internet Explorer that "autofills" the form field for you. You can tailor the AutoComplete feature to save and suggest only the information you want. You can choose whether to use AutoComplete for Web addresses, forms, and passwords, or not use it all. And you can clear the history for any of these.
This is what the Mac-version of the Explorer says: Turn Forms AutoComplete on or off:
|
|
|
SEARCH ENGINE POSITIONINGYahoo! increases pricesRecently search directory LookSmart added USD 100 to its Express Submit service. Now Yahoo! has gone and done the same. Yahoo! Express now costs USD 299! And no, you are not guaranteed a listing. The editors may perfectly well reject your site if they find it lacking. Still, a Yahoo! Listing remains one of the most efficient traffic generators on the Web, and we suppose companies will be willing to pay. Noncommercial sites may be listed for free. Read more about Yahoo! and LookSmart listings at Pandia (including a tip on a free backdoor into the LookSmart directory): http://www.pandia.com/sw-2001/51-submit.html AltaVista Trusted FeedAltaVista has also launched a new paid inclusion program for large sites, very similar to Inktomi Index Connect. Customers can determine which pages are included in the AltaVista index, and they can specify detailed page information including titles, descriptions and keywords. These elements are used to determine how the pages are ranked on the search result pages, as well as how the page listing appears. Partners provide information to AltaVista via an XML feed, which can be updated weekly. AltaVista provides detailed performance reporting for each URL submitted. AltaVista's Trusted Feed program is reserved for customers that submit more than 500 pages. Customers pay for each click-through. AltaVista Trusted Feed: http://www.altavista.com/sites/search/trustedfeed Intellectua will publish ebooks on search engine optimizationPandia editors Per and Susanne Koch have written two 3 minutes guides to search engine optimization for the Intellectua ebook company. "How to get good search engine rankings" and "How to get listed in Yahoo! and other search directories" will be published this fall. There will also be a separate ebook in Internet searching. http://www.intellectua.comFree Search Engine Optimization CourseDo you want to learn more about search engine optimization? In cooperation with 1st Search Engine Ranking.com Pandia can offer you a 5 day free email course on how you can get top rankings in the search engines. Send a blank email to 1stSearchRanking.1000.1072@optinpro.com to subscribe. |
|
BOOKSOnline magazineOK, we are not looking at a book this time, but a magazine for "information professionals". Information professional is a term covering anyone who has information retrieval as their job, including -- of course -- librarians. This is obviously a very specialized magazine, and it covers information that goes far beyond the needs of the regular searchers. Thus it not only covers regular search engines like Google and AltaVista, but also online fee-based databases containing articles, science papers, company information, public resources etc., as well as various types of electronic information products, CD-ROMs and search software included. And yes, the magazine does contain a lot of information on regular Internet searching. Greg Notess, the editor of SearchEngineShowdown.com, has his own column called the "Internet Search Engine Update". The main theme of the September issue was Web search engines. The magazine is highly recommended to all "serious" hunters and gatherers, and if you want a taste of what it has to offer, you can visit the magazine's website. It contains online versions of selected articles. Online online: http://www.onlinemag.net/index.html FINALLY...Do you like Pandia? Feel free to forward this newsletter to a friend. Click here to recommend the Pandia site to a friend: http://www.recommend-it.com/l.z.e?s=328530 Go to http://www.pandia.com/post/ to find information on how to subscribe and unsubscribe. The Pandia Post is edited by Per and Susanne Koch, Sign up for our free newsletter today! |
.
Pandia Post Home Page: