Calculating Fragment Matches with Xtend

In Author-it Xtend individual user settings are customized through the Options window.

User options let you set the level of similarity used by Author-it Xtend, controlling the number of text fragments returned as a match. Higher settings return fewer suggested fragments, but they should be a closer match to the text being typed. Lower settings will return a larger number of results, as the suggested fragments will have a lower match ratio.

Fuzzy matching is more exact so will return less results. Relevance matching is less exact so should return a larger number of suggestions.

Author-it Xtend uses a similarity matrix to make comparisons between text fragments. The results are based on the similarity between fragments, calculated by comparing the number of identical words in each paragraph.

• Fuzzy matching: looks at the similarity of the fragments and makes a comparison based on words common to both fragments and the order in which they appear in both fragments.
• Relevance matching: looks at the similarity of the fragments and makes a comparison based on words common to both fragments. Relevance matching does not take into account the order of the words.

Both methods make a comparison based on the unique words in a fragment, ignoring any subsequent instances of a word if they occur in the fragment. For example, in the fragment Install the sound driver after you install the video driver the words install and driver are only counted once.