Significant overhaul to file matching algorithm

As part of refining Alfred 1.2, This morning, I have re-looked at the way I am matching files in Alfred. For the longest time, Alfred has used a mix of continuous word based searching and wildcarded spaces. This has worked great in almost all situations but is weighted towards knowing the name of the file you are looking for.

For the purpose of this example, we are going to use the filename:

Boring Fiscal Report 2012 summary.txt

Previously, Alfred would find this if you typed words anchored and as they appear in order, e.g:

'boring fiscal summary
'boring 2012 txt

Alfred would also be happy if you wildcarded the first character, but this is a bit undiscoverable and certainly ‘techie’ e.g:

'*fiscal summary

This is far from ideal as the * isn’t the quickest key to get to, and you still need to know the order of the words in the filename.

New file matching for 1.2

Alfred will be much more intelligent in 1.2 by treating all words you type separately and matching files if all words exist in the file display name. This means that without knowing the exact filename, or needing to wildcard, you will be able to type:

'summary 2012
'2012 boring report
'txt report boring

… All of which will match this file.

This improvement is now in the latest 1.2 pre-release available from Alfred’s General > Update prefs :)


  1. preppeller posted this