in the context of information technology the key items are these, a population; what you are searcing through of which what you are looking for is called a "member", a target; The pupulation member that you want to find, a search devicee, process or engine; with what and how you will find the target if it is to be found, and a seearch index term or in common speech a keyword, but it can also be an image or a sound. I am sure that we have all had the experience of "That smell reminds me of...". The population is your stored memories, the engine is the process of recall and the index is the smell that remind you of ... The search index term is taken in and compared to a set of terms to see if there is a match and the engine then "fetches" whatever that matched term points to. That's just what is done, how it is done will be discussed later and if you've ever tried to write a search program, you know that it's harder than what you see here. The amazing thing about Google is that it can go through HUGE data banks in less than 2 seconds and find what you want or tell you that it can't find it. In fact doing it faster and better than everyone else is how Google made its name a household word "I'll just 'google [up]' such and such"

For 85% of what your are after. a simple search bar in your browser with AltaVista, Yahoo or Google will do perfectly and looking for a specialty search engine just isn't worth the effort. So keep 2 or 3 biggies in the hole. Mycroft will help you there.

Now, the other 15% is where the action is. this is the stuff you need to know in depth, wanto to know more about or just want to explore. THis is the job of the specialty search engine.

A search is a systematic organized hunt for something: Specifically, a binary subtractive process that separates out a smaller and smaller area until you find the smallest area in which your data is. For example, with just seven tries, I can find any specific number from 1 to 100 that you may be thinking of. I do it by taking the half-way point in what is available, in this case, 50, and ask if it is too high or too low. If I start at 50 and you tell me it's too high my next guess will be 25. If you tell me it's too low, I'll go with 75. Depending on the answer you give I'll take half of what's left. If 25 is too low, then I have between 26 and 50 so I'll say 37. If 75 is too high I'll try 62. In 7 tries I will have your answer unless you're being a smartass. Those half-way numbers I was guessing and getting closer and closer are called "hash"es and the rule I use is called a "hash algorithm", in this case a "binary search" since it's always one or the other (too high or too low). If I were tryig to guess a number between 1 and 10,000 I would need more tires. What a search engine does is has a table of data for which search term you plugged in that references a key or index. The table contains as data a set of terms that the engine tries to match. Thses terms has tags that the search engine can use to hash through saying "this tag is too high" or this tag is too low". when it finds the match, that term points and says "what you want is over there" and the engine, called a robot or spider, goes and fetches the material and brings it to you. Searching by an index is a gazillion times easier than going into each data file and saying "nope, nope...ah here it is" and takes probably a billionth of the tiem, too. Another kind of hash is rank in a group. I can say that so-and-so is somewnere between tenth and twentieth in rank, so I chuck out 1 through 10. This is called an "ordinal" algorithm". I can use "percentiles". If Tony Cobb hits in the 60^th Percentile, often writtne %ile, well he hits better than 59% of the guys in the league but not as good as 40% of them. Now if I wnat to find someone who hits in the 60^th %ile and see what the Runs Batted IN's and On Base % are and am looking for someone with 94 "ribbies" I plug in (usually to a professional and paid-for engine) 60 %ile and out comes Tony Cobb and maybe 30 other players with all the stats that they have. This is also done by dividing the population into quarters and using quartile or 1/4iles instead of %iles so that I can say that Jose De Cajones is in the second 1/4ile. If Tony Cobb hits in the 60^th %ile. has more than 95 RBI, an on-base % of better than 50 and I'm in th American league, I might want to make him my Designmated Hitter and "Plug him in the 4 hole". If his fielding and throwing stats are good, I may want to make himm a starter with a big fat contract and plug him in the 4 hole, unless I have to find the front of the horse as well, you know Manny Ramerez.

the process of using a search engine goes, from start to finish, like this

1. You've put the index term or keyword into your stock search engine and either go no results or the barest of minimum resutls that just is not enough but points to more and better.
2. You probably got back some reference to the topic of which this term is a part
3. You go to your search mill and find a meta-search engine. This will give you the power of several search engines all at onces. A true meta-search engine inputs the keyword as an index to several search engines at the same time, gets and presents results. Meta-search engines can harness over 20 search enginss at one time and some of these may be specialty search engines.
4. If you don't have what you need or just want more information, you look for and use a specialty search engine. These are operated by professional organizations or universities and access data on a specific topic like physiology, agriculture, sports or engineering that just would not be in a general or meta-search engine because it would not be relevent to enough people to be more than just clutter in a general search engine.
5. Among the tools in the well-stoked-up search mill are search engines to find search engines; usually topic-related to find specialty search engines. For instance, If I wanted to find detailed information about parmecia, I would go to a search engine for biology search engines, put in "protozoa" and "Paramecium". Among the returns I would get would be the names of search engines that specialize in protozoa (microscopic animals) that have data on parameicia

That is the basics of search science. The rest is practice.