Search
Search enables you to specify a word or a phrase to find relevant pieces of content without the use of navigation. 80% of IBM Documentation traffic comes in through search making it the primary entry point to IBM Documentation. IBM Documentation is part of IBM unified search and uses APIs to span its search across all technical content at IBM.
Scoped search
In IBM Documentation if you are on a product page, the site search is scoped to that product. If the search query can't find a match, it elevates the search to other versions of that product. If no search results are available within those versions, it then elevates to include all of IBM Documentation. This is done automatically, every time you search from a topic page. Searches triggered from family or product pages are scoped to that family or product. Search can be repeated for all content in IBM Documentation by clicking on 'all of IBM Documentation. If no results for a query string are found in the scope, the scope is expanded to all of IBM Documentation automatically.
Contextual search results
The search results include technical content from other IBM Tech Content sources by clicking the radio buttons in the other technical documentation sources on the search results page. Other technical content sources include IBM Developer, IBM Redbooks, Support TechNotes, IBM Training (formerly Skills Gateway), Announcement letters, and API Hub.
Search suggestions
When the IBM Documentation search engine finds 0 results for your search query, but it has a suggestion, it will search for that suggestion instead and prompt you with the suggestion.
Auto complete
When typing a partial word, after about 2-3 seconds, autocomplete suggestions appear as a dropdown box. When the partial word is contained within a product name then this product home pages are suggested as well.
Synonyms and abbreviation expansion
Abbreviations are expanded and search is executed with the abbreviation and expansion.
Spelling correction
The IBM Documentation search engine corrects typos in the search query. It performs the search with a suggested query that matches most to the original query.
Stop words
Certain words do not process within the search engine, mostly because they are too common or used as binding words. Examples of stop words include: ibm, was, etc. To search for WebSphere application server, use capitalized WAS.
Search results order
Results are ordered by score (higher on top). Each result has a score assigned to it based on the number of occurrences of the searched query in the Title and Body of the document. The score can be modified by other factors such as suggested matches.
In title search
Using keywords "intitle:" followed by your search term, or keyword, you can force IBM Docs search to scan only Titles of the topics. For example,
-
intitle:Help might yield the topic Help on Help
-
intitle:delimited might yield the topics "Delimeters" in the SPSS Statistics product docs
InuRL search
Use the inurl: operator to show results that contain the keyword you specify somewhere in the URL. This is useful if you want to find pages that are about specific topics. An InURL search query is always constructed as InURL: keyword. For example, "InURL: getting started" will return all URLs with the words "getting started" in the URL.
Read time
Each search result has a read time assigned to it based on the number of words in the topic and average human read speed.
- This functionality is for Product Documentation search results only.
- The read time calculation is based on an average read time of 200 words per minute / the number of total words. At this time, images and videos are not included in the calculation.
5-star rating
IBM Documentation search results page shows a 5-star rating on topics that have feedback collected from IBM Documentation's "Was this topic helpful" feedback mechanism. The 5-star rating calculation is derived from the feedback in the database. The rating calculation starts with using the feedback average (which is calculated by the total number of responses divided by the average negative and positive responses)
-
Average rating is calculated between -1.0 to 1.0 and all real numbers in between averaged to 2 decimal numbers using following calculation.
- net = pos - neg
- average = net / (pos + neg)
-
Stars are assigned based on the average (-1 = 0 starts – 0 = 3 stars – +1 = 5 stars … by 0.5 step)
Table of Contents filter
The Table of Contents (TOC) filter is in the TOC space on IBM Documentation left navigation section and is a way to further refine a search through the product documentation's table of contents. When you enter a keyword in the TOC filter, the filter crawls through your L1 and L2 titles and returns a list of all L1 and L2 topics containing that keyword or phrase. You also have the option of launching search from the bottom of the filter results.
Search tips
-
If one does not desire to match the substrings of the search query separately, then enclose the search query in quotes.
-
Using wildcard characters like "*" "." "?" does not bear any special meaning.