How can I conduct an effective search within the application?

Conducting an Effective Search within the Application

1. Understanding the search functionality

The CRM library search functionality provides basic and advanced search options. Basic search allows you to enter a search term and return all documents containing that word. This search only looks at key fields within documents such as title and description. The search returns a list of all documents containing your search term, allowing you to click on an individual document and view it. If a large number of documents are returned, it may be difficult to find the most relevant document. This is when the result sorting functionality becomes useful.

Search through the application is common functionality. It is often frustrating and unsuccessful. In the development module, the 'search functionality' adopts the same concept as internet search engines. The purpose is to return all documents and steps that are relevant to your search term. This differs from the simple search functions found in Document and Training management, which only look for document names and do not search through document contents.

2. Refining search criteria

This process can be repeated, and with current and new queries, the best search results can be displayed, and an option given for the user to store them in a folder for later reading. This process, while simple in description, is quite complex behind the scenes and can involve data mining and information retrieval techniques that are beyond the scope of this report. It is also not elegant to implement in terms of burdening the user with too much system interaction.

A simple initial method that can be employed for UI-based query refinement utilises the co-occurrence of keywords in the document collection to provide related search terms for any given query. For example, if we have a large database of digitised technical reports and a user performs a search on the term 'mobile agent', we can retrieve all the papers found and determine the frequency of any related keywords that were used by their authors. Taking the keyword 'mobile agent', it appears that one of the most pertinent papers to this user's query contains the phrase 'mobile agent migration'. The author may not have thought of this search and would simply use the search query again and get the same results. In such a case, we would like to inform the user that a query for 'mobile agent migration' may be a better one to meet their information needs.

Users are often not sure of the exact information that they want, and generating either too few or too many search results. For this reason, frequent keyword refinements are a common practice amounting to a complex search process, and one of the few search activities that are represented through more than one query. This is a powerful feature, but was not well supported in terms of the user interface in the old Citeseer. In particular, there was no history of past queries and it was difficult to combine old and new search results to help make decisions on what was the next best query. The query refinement functionality provided by Google is a good modern example for what Citeseer should strive for here.

3. Utilizing advanced search options

Step searching (Controlled Vocabulary): This enhances your search by searching for synonyms to a given word and it can be used in two different ways: - +: search for the word and all related words. For example, "+CAT" will search for "cat" and all words related to it. - ACRONYM: this will search for words beginning with an acronym. - -: turn step searching off to stop using any vocabulary changes made. For example, if you have previously turned it on.

Nested: This type of search uses brackets and works in a similar fashion to mathematical sums. It tells the system where to group the search. For example, you may want to return result A or B while C is a requirement. Without nested searching, this would be A OR B AND C.

Boolean: This uses operators such as "AND", "OR", "AND NOT", and "=". Here is an easy-to-read guide on how to use each operator correctly: - AND: narrows your search by only returning results that contain both words. - OR: will return results with either of the two words. This is good for when there are two words with similar meanings or when you are unsure about how to spell a word. - AND NOT: will exclude any results with the following word. - =: will search for the exact match for the word, which can be handy if the word is semi-relevant to what you are looking for. - Fuzzy Match: will return results of a similar match with differing characteristics. For example, "and" instead of "color".

Advanced searching can be used in your job to narrow your search and return the most relevant results possible. Below is a list of the types of advanced searches that you can use:

4. Reviewing search results effectively

There are a variety of steps you can take to meticulously review search results and find the records you need. Examine the search query expressed in the results. Larger-scale searches, both command-driven and keyword-driven, may result in a long list of candidate records. If you reached the results list from a large-scale search that has poorly met your information needs, you may need to reformulate your query and search in smaller increments to improve the precision of the results. For command-driven searches, the results list will typically indicate whether the search was performed on a primary or alternate access point. Evaluate the records in the results list to determine if they meet your information needs. If the search has retrieved an unmanageably large set of records, you may want to scan through a sorted display of the records to decide which are most relevant to your needs. If the result of the search is a small set of records, or a set of records deemed irrelevant or marginally relevant, you may need to broaden the scope of your search by using synonyms or truncation. This does not mean that you've overlooked a record that is just right but hard to find, or that the right record does not exist. Always stay flexible.

5. Troubleshooting search issues

After you've performed a search, utilize the highlighting feature to confirm the term was found in the search results. This will help the user determine if the correct search results were found and narrow the search if needed. To do this, select an individual record and navigate to the 'Find' menu and select 'Highlight Search'. This will highlight the search criteria in fields where it was found.

Unwanted search results occur when search criteria entered is too ambiguous or broad, resulting in too many results or no results at all. This may happen for a search on a particular form or for a specific record set. The search term may also be too specific resulting in no results found. For example, a user selects the Lease form and wants to search for all lease records with active status. If the user enters 'Lease' in the basic search field, the application will return all records from the Lease form. This is likely too many results. Instead, the user could go to the Lease form and do a search leaving all fields blank except selecting 'Active' for the lease status field. This will return more desired results.

If users are not obtaining desired results from the application search feature, more advanced tools and methods are available. This section will provide examples of searches with unwanted results and specify how to enhance these searches for a more desired outcome. Furthermore, this section will provide additional troubleshooting techniques to ensure end users receive optimal search results when using the IBM TRIRIGA application.