Introduction Structure of a pre-filter
index.html / bots.htm / termisearch.htm
Updated 25/MAI/2007, version 0.09 |
HOW TO BUILD A
A working example: Termisearch: English Italian French Spanish German
This approach is just a
proof of concept, hence, of course, in fieri.
You can modify the forms on this page in order to search for different targets: images, music, porn, books, solutions, confidential
documents, passwords or whatever.
With the Termisearch working example, below, I offer a linguistic (imho
quite interesting) approach just in order to demonstrate the utility and potential effectiveness of such simple filter tools.
the whole point of the "Termisearch" example below is to demonstrate how such "pre-filtering" approaches
can go beyond simple "webbit" results. Of course Thou does not need to go linguistic.
You will create your own pre-filtering forms at leisure using the schema below.
You may want to fine-tune special pre-filtering forms in order to search more effectively for books or images (eliminating
from the SERPs all those idiot sites that try to
'trap' searchers into advertisement hells or lists of crippled results for zombies).
Or maybe you want to retrieve MP3s or OGGs without having to wade knee-deep into morons trying "to sell" you those very
songs (quelle vulgarité!). Or whatever on such lines... I'm sure you get the infinite searching possibilities now in your own hands :-)
The Termisearch forms are experimental and must, should and can be further fine-tuned. Note that we can
also further fine-tune things there "subtracting words", thus eliminating less useful results that
often/always contain such words, see Termisearch's Spanish
The Termisearch example also shows that
drawback of using linguistic approaches is that this will work
quite well for languages that have great amounts of indexed pages/occurrences,
but won't give good results for 'smaller' languages.
So, does the "web at
large" deliver better terminological quality (and quicker and to no cost) than
trusted tools and dictionaries? Of course not, or at least not always. But
the 'termisearch' small tool below may still prove useful for linguists. Of course the results
will vary depending of the 'notoriety' and 'trendiness' of your query... such are
the tides of the Web.
|Structure of a pre-filter |
See the complete code of all the various Termisearch forms below (with Opera: CTRL+F3), here just an
example taken from
the Spanish termisearch form below.
<form target="_blank" name="gs" action="http://www.google.com/search" onsubmit="return ShowFormWarning()" method="get"> What and where we do
<input size="30" name="q"> Width of the input field
<input type="hidden" value="100" name="num"> How many results per page we want
<input type="hidden" value="off" name="safe"> No puritanical crap
<input type="hidden" value="en" name="hl"> Englisz for broadness
<input type="hidden" name="lr"> Parameter 'language restrict' ignored here
<input type="hidden" value="ISO-8859-1" name="ie"> Input Encoding - this parameter is deprecated and
ignored. Use UTF-8 encoding instead unless you know what you are doing.
<input type="hidden" value="en.espanol se.dice denominado denominada se.usa que.significa" name="as_oq"> as_oq
The real important ORRED terms
<input type="hidden" value="libros.en.espanol" name="as_eq"> as_eq
: What we do not want
<input class="submit" type="submit" value="Button_text" name="btnG"> Gimme a nice button text for submitting
<input type="hidden" value="0" name="filter"> Uh? "repeat the search with the omitted results included?" Yep, please
So, that's the theory, now, as an unrelated example, look at the following banal 'book searching' mask:
BOOK SEARCHING (1) (2) (3)
"rapidshare OR megaupload OR yousendit OR filefactory
"NOT amazon NOT sex NOT tits NOT win"
1) Change the book title according to your taste and inclination, duh
2) Always use the CACHED copies of google's results to bypass simple locks, duh
3) Exempli gratia, try inputting java.in.a.nutshell or more simply in.a.nutshell to check spreads
However, as said, this pre-filtering approach can deliver results that go
waay beyond such 'webbit similar' book findings, as my
'termisearch' example below will (I hope) demonstrate.
|The 'termisearch' example |
The raison d'être of the very simple forms given on the Termisearch working example is to find
definitions, possible translations and/or explanations of a given term IN ANOTHER
So, for instance, to check how "offshore.outsourcing" has been/is usually translated INTO GERMAN (or INTO FRENCH, or
INTO ITALIAN, etc.) Or, the other way round, how "Weltanschauung" or, say,
interinale" has been/is
usually translated INTO ENGLISH.
Or the real meaning of a proverb... or whatever on these lines.
This is below simply done
adding, subtracting or ORring
automatically some ad hoc search terms to whatever query you may have.
quotes around your searching terms for exact match (or join them through a
Therefore: "offshore outsourcing" or offshore.outsourcing, else you'll gather pages with only offshore and/or with only outsourcing
OR translated OR means OR in.english"
OR vuol.dire OR che.significa OR vuole.dire OR in.italiano OR
littéralement OR signifie OR veut.dire"
OR se.dice OR denominado OR denominada OR se.usa OR
OR heisst OR auf.deutsch OR heisst.auf.deutsch"
(c) III Millennium: [fravia+], all rights reserved, reversed, reviled and revealed