TIPS ON SEARCHING
The search engine on PALawHELP.org uses a search engine that is a natural language, full-text search that supports both simple word searches and phrase searches. The information below introduces you to how to search this site, and covers:
- How does natural language searching work?
- Which content items does the search engine include?
- Which parts of each content item does the search engine examine?
- How are search results displayed?
- Can searches be "refined" to improve the results?
- How often does the search index update?
How does natural language searching work?
The search engine is designed to perform searches based on natural language - the way people talk and write phrases. This means that users are able to enter their searches using common spoken or written language. There are no complicated syntax rules or search operators to remember.
Searches also automatically include common variations on the main terms used in your search. For example, a search for "divorcing" will match resources containing the word "divorce." Similarly, a search for "what is elder abuse" will match resources containing those search terms as well as related words such as "abusive" and "elderly."
Currently the search engine supports natural language searching in English and Spanish.
The search engine also contains an index of common synonyms to support natural language processing and make it more likely the search results will find what you are looking for.
For example, a search for "dissolution" will match resources containing the word "dissolution" as well as resources containing "divorce."
The search facility also has a suggested spelling feature for commonly misspelled words. For example, a search for "afadavit" will result in the message "Did you mean to search for affidavit?"
Which content items does the search engine include?
The search engine looks at all current Resources, Organizations and News articles appearing in PALawHELP.org. Expired news articles and other materials not currently active on the site, such as materials which are being updated, are not included in the search results.
Which parts of each content item does the search engine examine?
The search extends to all of the major content areas of the site.
When examining Resources, the search engine includes the following fields:
- Resource Title
- Resource Author (organization and Other)
- Resource Description
- All text in resource attachments (including PDF documents, as long as they've been scanned "OCR," which converts image to text)
- If the resource is an FAQ, all text and link fields are also included.
- If the resource is a Short Profile, all description and name fields are included.
When examining News items, the search engine includes the following fields:
- Title
- Teaser
- Body Text
- Author/Source
When examining Organizations, the search engine includes the following fields:
How are search results displayed?
Resources located by the search are displayed first, ranked in order of relevance. Relevance is determined by a number of factors, including in which field(s) and how often the keyword(s) appear.
For example, if a keyword matches a word in the resource title, that item will rank higher than an item with the word matching in the resource body text only. Similarly, the more frequently a search term occurs in a document, the higher it will rank.
Resources have been manually "boosted" in the search results and will generally rank higher than News and Organizations. However, an Organization may be ordered above Resources and News items if it is a strong match.
Best Match is the default order in which results are displayed. Selecting Newest orders the results by date created with the most recent at the top of the results
Using the optioins in the right navigation bar, you can filter search results according to:
- Content type (Resource or Organization)
- Channel /Category
- Topic / Subtopic
- Language
The number of matching items displays in parentheses next to the filter.
Can searches be "refined" to improve the results?
You can make your searches more specific by combining the words you used for basic searches with characters such as quotes and plus signs.
- " " (double quotes): Use quotes to make sure that only the specific word or phrase within the quotes is found.
Example: "bank" will only find the word bank and not banking, banked or banker. Note: If there is an uneven amount of quotes in a search, the quotes will be ignored all together.
- '+' and '-' characters: '+' and '-' characters are treated as "mandatory" and "prohibited" commands for the words that come after them.
Example (+): To see items that only have Legal and Services in the item, search Legal + Services Example (-): To see items that have the word landlord without the word tenant in the same item, search landlord - tenant
How often does the search index update?
The index is updated shortly after any additions or changes are made to Resources, News, or Organizations. This indexing process happens in the background with additions and changes being reflected in the search results within five minutes of their being made.
How are non-English resources searched and returned?
Non-English resources are searched the same way English language resources are. However, if the search is conducted from the English site, the search engine returns the English resource as the primary result, with the alternate language version(s) linked below.
Note: The search engine currently only supports searches in languages that use the Latin (aka Roman) alphabet. The search cannot match search terms in languages that use different alphabets.
For example, if you enter a search using Russian characters, the search cannot match Russian resources containing that text. For more information on what languages are supported, please email support@probono.net. |