[Previous]
[Next]
[Up]
[Title]
6.2. Search methods
Two
types of search/display options are provided at this stage.
Simple Search 131-is the search method recommended for
inexperienced users, and is the default option for DIAL Search. Users can enter
text in any form (eg `I want laws on tax and bankruptcy' or a few key terms
like `tax taxation bankrupt bankruptcy') without knowing about search
connectors. Results are ranked according to likely relevance. The search
results will sometimes lack the precision that an Advanced Search (which uses
boolean and proximity connectors) can provide, but for most purposes is safest
for inexperienced users (assumed to be the main audience of Project DIAL).
Also, since search connectors and searches for phrases cannot be used in Simple
Searches, and use of truncation can distort results, it will be insufficient
for more experienced users.
Advanced Search 132 allows searches using boolean and
proximity connectors (and requires them except where a search is for a single
word or phrase). These results are also ranked in likely order of relevance.
This combination of boolean/proximity searching and relevance ranking allows
reasonably broad searches (to aid completeness) with the relevance ranking then
providing more precision.
As a general rule of thumb, Simple Searches are usually satisfactory when each
concept that you are looking for can be represented as a single word, or you
are searching for some unusual name of a person or thing. Otherwise, Advanced
Search is probably required.
It will often be sensible to try a Simple Search first, and if the results are
not completely satisfactory, try an Advanced Search.
The following guidelines for users (derived from the Guided Tour materials)
explain the effective use of the Simple Search option:
- Enter one or two key words to describe each concept relevant to your
search, in any order. Example:
- To search for laws on pollution of water resources, enter `pollution
water'. These two key words represent adequately the key concepts of the
search. Synonyms could also be used if it was felt that `water' might not
always be used, resulting in a search such as `pollution water river'
- Truncation of search terms using `*' is allowed, but sometimes distorts
the relevance ranking of results because each variant of the truncated term is
regarded as a separate significant term. Example:
- To search for laws on taxation in India, enter `tax taxation India' and
most of the top-ranked results will concern India. A search for `tax* India'
will not be effective because the numerous words starting with 'tax' will
'drown' the reference to India in the relevance ranking, and almost all of the
top-ranked results have nothing to do with India.
- If you use any search connectors they are ignored. `Tax near bankrupt' is
the same as searching for `tax bankrupt'.
- All commonly occurring words are ignored (eg but, the, an, of, not, in,
and, or).
- It is impossible to search for a phrase as such. You can enter a phrase or
sentence as search terms, but the result will be the same as searching for the
words in any order. Example:
- A search for `copyright in sound recordings or music' can be entered but
the words 'in' and 'or' will be ignored, and 'sound recordings' will be treated
as two words, not a phrase. The search is the same as `sound music copyright
recording'. The result is nevertheless satisfactory, as shown below. The
top-ranked items were major pieces of copyright legislation from India,
Australia, Vietnam, Zambia, New Zealand and China.
First 10 of 2253 results for search `copyright in sound recordings or music'
The following guidelines for users (derived from the Guided Tour materials)
explain the effective use of the Advanced Search option:
- Search connectors must be used if more than one search term is used
(unless the words are part of a phrase). The boolean and proximity connectors
which may be used are listed below.
- Truncation of words using `*' increases completeness and has a less
distorting effect on relevance ranking because the use of boolean and proximity
connectors eliminates completely irrelevant instances before ranking takes
place. Examples of desirable truncation are:
- `negligen*' for 'negligent, negligently, negligence (etc)';
- `Viet*' for 'Vietnam Viet Nam, Vietnamese'
- `pollut*' for 'pollution, polluted (etc)'
- Synonyms should be used, but `or' must be used to separate them (eg
`weapon or gun or pistol', `privatization or privatisation').
- Phrases are recognised automatically as phrases and do not need to be put
in quotes. eg freedom of information.
- Regular singulars and plurals are automatically found (eg `bank' finds
`banks' ) so truncation should not be used merely to find regular plurals (eg
`bank*' is not needed to find `banks').
The
following connectors may be used:
- AND - two words or phrases on the same page. Examples: tax and
bankrupt; Canada and privatization
- OR - either of two words or phrases on the same page. Examples:
`trademark or trade mark'. A Simple Search would not be so effective, because
'trade' and 'mark' would be treated as separate, unrelated words
- NEAR - two words within 50 words of each other. Examples: `pollut*
near river' or `(privatization or privatisation) near (power or electricity or
gas)'
- W/n - two words within a specified number (`n') words of each other
. Examples: 'jud* w/10 salar*' means 'judge, judicial etc within 10 words of
salary or salaries'
Parentheses should be used to ensure that connectors
operate how the user intends them to. For example `(river or catchment) near
pollut*' will give the desired result, but `river or catchment near pollut*'.
An effective rule of thumb to prevent errors is that parentheses should always
be used wherever a search contains more than one type of connector.
[Previous]
[Next]
[Up]
[Title]