InfoJukebox represents a new paradigm in Web searching.


Instead of locating pairs of words that appear on the same page...

Welcome to Bob's Turtle and Snail Page ... We hope you've enjoyed your visit.

...InfoJukebox searches for words that appear in the same thought.

He was the fifth of seven sons of William Henry Hope, a stone mason, and Avis Townes Hope, a former concert singer. When he was four, his father brought his family to Cleveland, Ohio. On December 20, 1920, by virtue of his father's naturalization, Bob - the name by which the world would later know him - and his six brothers also became U.S. citizens.

Other search engines present you with pages which may or may not fulfill your requirements, and leave you to open them (if you can!), then search them for keywords -- using primitive techniques like "Find (on This Page)". InfoJukebox immediately shows you why a page has been accepted. Only pages which have been opened and validated will appear.

In addition, like the rest of our site, InfoJukebox is completely interactive. Thanks to the modern workday miracle of DHTML, you can pick and choose from the selections that InfoJukebox originally makes. You can drag from "Accepted" to "Rejected" frames, click-and-clear, save URL lists to your private Play List, combine Play Lists, even export to Excel with a single click.

The cost for this service is zippo. (Just download our free InfoJukebox control.) And when we say "free", we don't mean you don't pay, but we collect lots of information about you that we sell to marketers. Let's be absolutely clear about this. Nothing you do, search for, save, store, or think about while using InfoJukebox gets seen by anyone but you (and of course, the search engines themselves). We do not receive any information from you or your computer. We don't know who you are. If you want to contact us, you'll have to use email!



One Hundred and One Uses

Why should you use InfoJukebox, when there at least one or two other search utilities on the Internet? Let's take Software Development, as an example. Let's say you're developing a Web utility which uses Microsoft's Internet controls (as, coincidentally, we were recently). Say you look up the AsyncRead method in Microsoft's otherwise wonderful Documentation Search Wizard. And you read the following:

vbAsyncTypeByteArray The data is provided as a byte array that contains the retrieved data. It is assumed that the component author knows how to handle the data.

...but you don't have the foggiest idea.

Well, the world's a big place. Somebody must know something about this topic, and not be so coy as Microsoft about parting with the information.

Using InfoJukeBox, you can type in: AsyncRead vbAsyncTypeByteArray
and let'er rip. What comes back in a few minutes is:

http://www.miken.com/htmlzap/hzdoc.htm
AsyncRead "http://poit.narf.org/brain/index.htm", vbAsyncTypeByteArray you'll need an event handler that looks something like this (assuming your HtmlZap control is called "HZ"): Private Sub UserControl_AsyncReadComplete(AsyncProp As AsyncProperty) Last modified: 25-Jul-97 14K (AV) (NL)
Title: HtmlZap Documentation


http://deutsche.devx.com/0597vbpj/RZ0597/rz0597.htm
Vielleicht haben Sie sich über die Konstante vbAsyncTypeByteArray im AsyncRead-Aufruf gewundert. In unserem Fall, in dem Sie einen Textstring anfordern, ist Ihre Überraschung durchaus berechtigt. AsyncRead kann jedoch drei Anfragetypen handhaben: Dateien, Byte-Arrays und Bilder. Wenn Sie eine Zeichenkette empfangen möchten, müssen Sie die URL als ein Byte-Array anfordern, und dieses Array mit der Funktion StrConv in eine Zeichenkette konvertieren. Die beiden anderen Konstanten sind vbAsyncTypeFile zum Empfang einer Datei und vbAsyncTypePicture zum Empfang eines Bildobjekts.
Last modified: ? 30K (ISK)

The first one's in English and it looks promising, but they actually don't say anything much useful on this topic. The second one's in German, but it's worth a look, hey? And a few pages down, we find this:

Private Sub UserControl_AsyncReadComplete(AsyncProp As AsyncProperty)
On Error Resume Next
      RaiseEvent DocReceived(StrConv(AsyncProp.Value, vbUnicode),AsyncProp.PropertyName)
      If Err Then
            CancelURL AsyncProp.PropertyName
            RaiseEvent DocReceived(Err.Description, "Failure")
      End If
End Sub

The point is, we get the information we need, without manually scouring a vast wasteland of inapplicable pages. (If you're interested, the tidbit Microsoft assumed we'd know is that ByteArrays need to be converted back into strings using the StrConv(AsyncProp.Value, vbUnicode) command.
   


InfoJukebox Feature Guide

InfoJuke box is so different from other search engines, it's worth a few minutes of your time to review its features.

First thing you'll probably do, is sit down and type in your own name:

"John Z. Smith"

If you do this, you are correctly using quote marks. But you're not really utilizing the full potential of InfoJukebox, since its main strength is to bring back juxtapositions of two or more concepts (words or phrases). On the other hand, this is a quick way of running this simple search past the various search engines we've included in the InfoJukebox search drop-down list.

A more advanced use of InfoJukebox would be:
"John Z. Smith" "Master builder"

Note that "+" marks are not needed. InfoJukebox assumes that you're looking for both of these terms. This always holds true, unless you use the "or" keyword. For example, you might want to find any occurrence of "John Z. Smith" paired with either "Master builder" or "Architect". To do this, you could type:
"John Z. Smith" ("Master builder" or architect)

This tells InfoJukebox the you really want to see "John Z. Smith" in the citation (don't come back without it), but you'll accept either "Master builder" or architect. As soon as you click Search, InfoJukebox sets the "at least two words" flag, because if you insist on "John Z. Smith", but don't insist on either of the other two words/phrases, that implies you'll accept a citation that doesn't contain one of them.

You could also spell this out more explicitly by typing:
"John Z. Smith" and ("Master builder" or architect)

Note that InfoJukebox tries to limit the number of search engine calls it has to make, so even though a technically correct approach might be to call, for instance, Alta Vista with the two search strings...
+"John Z. Smith" +"Master builder"
+"John Z. Smith" +architect

...InfoJukebox doesn't always do this. For search engines like Alta Vista, which support the "+" operator, we take advantage of this ability and use the "at least N words" method. Instead we would pass:
+"John Z. Smith" "Master builder" architect

If you want to override this behavior, and force InfoJukebox to the stricter multiple-call technique instead, click the Options keypad search terms box to get to the Search Strategy frame, and set the Strict interpretation of "or" checkbox.

As an example of where multiple searches become necessary (no matter which search engine is used), let's say we're trying to find out about the children of Mr. Smith and Ms. Jones. We could enter the following:
(Smith or Jones) and (son or daughter)

Now InfoJukebox will make two calls to each search engine being used:
+Smith (son or daughter) (at least 2 words)
+Jones (son or daughter) (at least 2 words)
This will yield back citations that contain either (a) "Smith" and "son" (b) "Smith" and "daughter" (c) "Jones" and "son" (d) "Jones" and "daughter"

For Google and other engines which don't support "+", a total of four separate searches will be performed. The four searches (a)-(d) will also always be applied if the Strict interpretation of "or" checkbox is set to ON, as noted above.

Currently, InfoJukebox cannot handle nested parentheses.


Well, all that's been the easy part. Now it gets interesting.

Up to now -- that is, with other search engines -- when you typed a plus or minus sign, you were saying, "only include pages" or "don't include pages" that have this search term. But InfoJukebox is working on a finer scale. So the plus and minus symbols mean something different. They mean, "only include citations" or "don't include citations" that have this search term. (A citation is the block of text that shows why the page was accepted.) In addition, you might also want to include or exclude pages that contain particular terms. For this, we've substituted the "*" and "/" symbols.

Here's an example. Let's say we want to find references to the wonders of modern science. We could try the following search string:
(wonder or marvel or miracle) science

This does a pretty good job; again, InfoJukebox sets the "at least two words" flag, because "science" is required but other words aren't. (You can actually overrule this decision and change it to "at least 3 words". That would mean you insist on the word "science" but will accept two or more of any of the other words.)

http://www.geocities.com/CapitolHill/Lobby/8765/marvelormiracle.html
   Marvel of Modern Science or Miracle Lets assume that there are in excess of six billion humans inhabiting planet earth. Now, let's pretend (with everything else being equal) that on a given day (e.g.; January 7, which is my birthday) not a single baby is born on earth .

But let's say that during this search we see that there are an awful lot of citations having to do with science-fiction ("Wonder Stories, Science-Fiction" eg.). We might do better by requesting that "fiction" not appear in the citation. What's more, we notice that there's a lot written on the subject of miracles by folks who are more interested in the religious aspects than the scientific. Often, they contain the word science, but it's not exactly the same "science" we're thinking of.

Well, I'm sure the creation-science folks will excuse us if we just go ahead and rewrite the search string as follows:

(wonder or marvel or miracle) +science -fiction /religion

This yields back only citations containing one of the three words "wonder", "marvel", "miracle" in addition to the word "science" -- of course, without the word "fiction" appearing anywhere in the citation itself. And although you won't see this at first glance, the citations themselves now come only from pages which do not, ever, mention the word "religion".




Some notes on the March '00 release.

New features include:

The ability to combine "*" and "/" operators with "and" & "or". Here's a variation on the example above. Let's say we first change the search criterion to:
(Smith and son) or (Jones and daughter)
Instead of the four combinations cited above, this concentrates on only two. Now let's restrict it even further by adding a page restriction:
(Smith and son) or (Jones and daughter) *politics
This will search for the two combinations (Smith+son,Jones+daughter) but only return results from pages that contain the word "politics".

What's more, you can include "*" or "/" operators in the individual branches of the search.
(Smith and son /Democrat) or (Jones and daughter *Republican) *politics
In this case, the same two combinations would be returned, but those that met the requirements by virtue of mentioning "Smith" and "son" would only be accepted from pages that did not mention the word "Democrat" (at the same time as it did mention "politics")! Citations containing "Jones" and "daughter" would be accepted only from pages that mentioned both "Republican" and "politics".