Smart Computing Article

Subscribe Today | Contact Us | Register Now

Home | Tech Support | Q&A Board | Article Search | Subscribe & Shop

Bots

Email This

Print This
View My Personal Library

Web Sites March 2000 • Vol.6 Issue 3	Add To My Personal Library
Bots Shopping, Finance & Chatter Bots Learn To Search For You

The dawn of the bots in cyberspace always promised to bring fruitful days to Web users. It's just taken longer than many people expected for the sun to rise.

The original vision of bot technology on the Web, that of silent herds of tireless task-hounds combing the furthest reaches of the Web for pearls of information, gave way rather unceremoniously to a somewhat less-inspiring reality: hordes of drones, misdirected by information overload, wandering in endless search loops and clogging regular Internet traffic along the way.

But the promise remains, and if the mid-90s represented something of a dark age on the Web, bots seem to be emerging into a Renaissance. Today's bots come in all shapes and sizes, and they perform all manner of tasks. You can employ a spider bot, for example, to scour the Web for documents containing information about a certain topic. A stock bot can retrieve investment information for each new IPO that hits the market. Webmasters use bots to maintain links and create indexes for their sites. Marketing companies use bots to gather information about consumer demand. And a good e-mail bot will automatically can the spam you receive from marketing companies who unleashed their research bots on you.

In this article, we'll survey the current crop of Internet bots, examine a few popular representatives, and look ahead to the not-so-distant future of bot technology.

What's Hot.

Hands down, the leaders of the bot revolution, at least in terms of popularity among Web surfers, are the shopper bots. Given the massive groundswell of online retailing in the past year or two, the popularity of these bargain-savvy critters stands to reason. Retail competition increases with each new merchant that sets up shop on the Web, and the price-conscious online consumer can spend hours looking for the best deal possible on that new pair of rollerblades. Shopper bots can eliminate a huge amount of browser time by doing virtually all of the hunting for you.

While the best shopper bots can search a vast number of online retail sites for a department store cornucopia of merchandise items, be aware that even the developers of shopper bots are hoping to make a buck or two. This is the reason many comparison shopping sites have forged cooperative deals with specific online merchants. A conflict of interests? Not in today's endlessly cross-marketed economy. And the overall usefulness of such bots is still ample. Still, in accordance with the old adage, caveat emptor, it might pay to use a couple different shopper bots in tandem.

The following are a few of the more popular shopper bots currently awaiting command to do your bidding.

DealTime (http://www.dealtime.com) lets you search for products based on item, price, manufacturer, and even model number. Because you can define a search period by entering a date range, the DealTime bot can keep hunting long after you've decided to pack it in, keeping you informed of prices and locations via e-mail, your personal pager, or a specialized icon you can download to your desktop.

Auction hounds can use BidFind (http://www.bidfind.com) to scour more than 300 online auction sites for everything from antiques to left-handed golf clubs. And if you can't find the item you're looking for, you can post it on BidFind's Wanted page for all potential sellers to see.

The bots of the Jango/Excite Product Finder (http://www.jango.com) crawl their way through most categories of merchandise, sifting through merchant sites, auction sites, and classified ads. If you just can't seem to choose between competing brand names, you can use Jango's review finder to scare up product reviews for the item in question.

R CLEAR=ALL> The mySimon bot "studies" your buying habits so he can scour the Web for the best deals.

The mySimon bot (http://www.mysimon.com) uses a proprietary technology called the Virtual Learning Agent to mimic the common shopping patterns of the average consumer. Simon, the bot's lantern-jawed mascot, retrieves price, shipping, and warranty information from thousands of online merchants selling everything you could find at the mall, and probably more.

As a category, shopping bots are rapidly evolving into multifaceted consumer services. Companies, such as San Francisco-based Active Research (http://www.activeresearch.com), are integrating shopper bots into full-blown recommendation engines. Recommendation engines don't just compare prices, but they also compare value, based on personal preferences that you enter and shopping habits you exhibit. Such megabots can do more than just make buying recommendations based on your consumer profile. In theory, they'll be able to actually preempt your purchasing whims, say, by notifying you of the availability of that priced-to-sell 42" Sony Trinitron you know you'd love. To many consumers, this concept may seem as irresistible as it does frightening, which is why we think integrated recommendation engines have the potential to revolutionize the online shopping experience before next holiday season. At press time, for example, Lycos, GoNetwork, and mySimon all had signed on to incorporate Active Research's flagship product, the Active Buyer's Guide, in their search pages.

On the heels of the shopper bots are finance bots, which can gather personalized investment information for the market novice and the intrepid day trader alike.

For example, you can retrieve company information and track performance statistics for any group of corporations you choose using the CompanySleuth bot (http://www.companysleuth.com). The Navigate One bot retrieves Web links to all financial information associated with a stock symbol or currency you enter. You can access a free Quick Search version of NavigateOne (http://www.NavigateOne.com/quick/Quick.asp), or you can register for a free trial period of the bot's professional version (http://www.NavigateOne.com/quick/Upgrade.asp). The Finance Wise search engine (http://www.financewise.com) dispatches its bots only to large but select groups of sites that are making headlines in the high-stakes financial world, collecting focused research and performance data for just about every money topic imaginable. And for $9.95 per month, the IPOPatrol bot (http://www.ipopatrol.com) monitors the IPO notification pages E*TRADE, DLJdirect, and Wit Capital, apprising you with an e-mail of every new Initial Public Offering it sniffs out.

The specific bots discussed in this article represent a few examples of the many bots currently in circulation. For a comprehensive listing of all known bots in all categories, with links and summary reviews, check out the BotSpot (http://www.botspot.com).

What's Not.

The days of the generic search engine, the basic keyword locator that functioned the same for everyone, are long over. Virtually all major online search sites now cater, at least to some degree, to the strong desire among users for personalized service. Lycos, Excite, DealTime, and mySimon, to name just a few of many examples, all let you customize a personal start page containing continually-updated links to Web content you specify. Many sites send you an e-mail or notify you by some other method you specify when news breaks in your areas of interest. Today's bots also cater to the "virtual community" in new ways. Lycos, for example, uses intelligent agent technology supplied by WiseWire (http://www.wisewire.com) to "learn" from community opinion as it assembles Web content into topic-based channels, or wires.

In general, however, in the current field of bot development there seem to be few truly irrelevant areas. One can always judge a technology by its popularity with users (Betamax VCRs come to mind). But at its core, bot development, unlike Web browser enhancements or new releases of office software suites, is inherently about discovery. So while a shopper bot will always be more widely used than a specialized intelligent data-mining bot or a bot designed specifically to help programmers create other bots, the former technology isn't necessarily more important than the latter.

Still, some classes of bots are further from fulfilling their true potential than others. Chatter bots, for example, are popping up around the Web in increasing numbers, but most are still functionally equivalent to novelty items.

Chat bots use natural language analysis to conduct semihuman conversations with users. Some chat bots are especially clever about disguising their robotic natures, intentionally using eerily human techniques, such as typos or bad grammar. The current crop of chat bots is primarily useful as entertaining, high-tech Mad Lib games. For example, if you strike up a conversation with Erin, the chat bot bartender at Extempo's Virtual Bar (http://www.extempo.com/webbar/index.html), you're likely to get some snappy comebacks and an entertaining dose of Rock �n Roll trivia. ALICE the chat bot (http://206.184.206.210) uses words and phrases she has learned from conversations with people all over the world, and she can generally maintain a semirealistic conversation of up to three or four exchanges before breaking down.

R CLEAR=ALL> Alice the chatter bot "learns" to converse from the questions users ask.

The most important thing about chatter bots is their future potential, which seems endless, depending on the rate of advancement in the field of Artificial Intelligence (AI). Automated help desks run by intelligent chat bots, for example, are soon to be commonplace. Travel agent chat bots will help you plan trips. The software developer Neuro-Media(http://www.neuromedia.com) already employs a chatter bot named Shallow Red, a tongue-in-cheek reference to the infamous IBM supercomputer Deep Blue, to answer common questions about the company and its products. These examples represent only the tip of the AIceberg where bots that speak like humans are concerned.

What's Next.

Without question, the future of bot technology will be defined by the continued development of AI.

AI is the branch of computer science concerned with developing computer systems that can think, reason, learn, and act without direct human intervention. In theory, a fully intelligent bot would be able to perceive its own environment and modify its behavior in response to it. A fully intelligent bot would carry out your requests, learn on its own, handle increasingly complex tasks through independent reasoning, and even anticipate your needs based on what it "knows" or has recently "learned."

Currently, this level of AI is still the stuff of which science fiction is made, at least as far as mainstream Internet users are concerned. But thanks to the flurry of development that has occurred over the past several years, each new generation of bots is smarter than the last. Many data mining bots already use their own internal logic to sift data and refine their search methods as they go; we believe that this class of bot, perhaps more than any other, will be integral to distilling actual information from the ever-growing mass of data on the Internet.

But MIT's Things That Think consortium (http://www.media.mit.edu/ttt), a group of top computer scientists devoted to the field of AI research, believes the future of intelligent bots is hardly confined to a computer screen. "Your shoes," says the TTT consortium, "should be retrieving the day's personalized news from the carpet before you even have time to take off your coat. We must expect more from our environment."

At least we can await the coming of spring with our bare feet on the numbskull throw rug while we wait for the footware that knows more about current events than we do.

by Sean Doolittle

What's HOT	What's NOT	What's NEXT
�Shopping bots �Finance bots �Personalized bot services	�One-size-fits-all bots �Keyword locators �Novelty chatterbots	�Recommendation engines �Artificial intelligence �Intelligent chatter bots and data-mining bots

Terms To Know

agent�Common usage has made the terms bot and agent essentially interchangeable. By definition, an agent is a piece of software that can carry out a specific mission without direct human intervention. The mission of a software agent usually consists of automating a task or retrieving some piece of information and bringing it back to the user.

artificial intelligence (AI)�AI is the branch of computer science concerned with instilling humanlike thought, logic, and learning capabilities in computers. MIT professor John McCarthy coined the term artificial intelligence in 1956. Today, the field of computer AI has branched into games and entertainment, decision analysis, natural-language interaction, neural networking, and robotics (both hardware and software). True artificial intelligence has not yet been achieved; however, developmental advances in AI are visible, particularly in the field of computer games. Advanced computer chess programs, for example, have beaten humans, most notably when an IBM supercomputer known as Deep Blue defeated grand master Gary Kasparov in 1997.

bot�Short for robot. The Merriam-Webster's Collegiate Dictionary defines robot as "a device that automatically performs complicated often repetitive tasks." A bot is a software robot that automatically performs some computer task. MIT professor Joseph Weizen-baum invented one of the first working software robots in 1966; named Eliza, this program notoriously mimicked a psychotherapist by answering user questions with questions of its own. Eliza worked by analyzing the language of the questions asked, then matching identifiable linguistic patterns with preprogrammed responses.

Many modern software applications employ bots to simplify tasks. The Hypertext Markup Language (HTML) development package Microsoft FrontPage, for example, offers an add-in component containing bots that automate many tedious aspects of Web page development, such as displaying the date and time.

The bots discussed in this article are Web-based programs designed to automatically perform tasks, such as finding and retrieving information. Bots can be highly specialized, and some bots use internal logic to "learn" from experience.

data mining�The process of examining large masses of information for specific characteristics, trends, or patterns. For example, data mining bots can search online databases to help marketing research companies find users with common surfing habits. True data mining doesn't just present existing data in alternative ways; it actually discovers new relationships among data.

search engine�A search engine is a program that searches documents for a specific keyword or phrase, indexes what it finds, and returns a list of locations where the documents reside. A search engine works by sending out a specialized bot called a spider bot to retrieve as many documents as possible. AltaVista (http://www.altavista.com), Yahoo! (http://www.yahoo.com), Lycos (http://www.lycos.com), and Excite (http://www.excite.com) are among the most popular engines for searching the World Wide Web.

spider�A bot that automatically retrieves Web pages and delivers them to a search engine. A spider "crawls" the Web in search of documents, bringing what it finds back to the search engine's database. Hence, spiders are sometimes referred to as webcrawlers. Metasearch engines, such as MetaCrawler (http://www.metacrawler.com) send multiple spiders to other search engines to retrieve and catalogue information that other spiders have already found. Large search engines, such as AltaVista, have many spiders working together.

Want more information about a topic you found of interest while reading this article? Type a word or phrase that identifies the topic and click "Search" to find relevant articles from within our editorial database.

Home Copyright & Legal Information Privacy Policy Site Map Contact Us