I believe a 'list of equivalents' or synonym list would be a great thing to consider for InfoQube.
Reason for suggestion
When I am in meetings etc I take many notes (as most people will). I use varying terms to refer to certain things- for example I'm running a "Transactional Banking" project which is referred to in my notes as "Transactional Banking", "Trans Banking", "TB" or "Banking Transition".
Same goes for references to people. References to Shane North in quick notes could be: SN, Shane N, S North, Shane North, Shane or even 'Credit Manager'.
Implementation
Pretty simple I'd hope for the benefits it would add.
Basically a listbox, each line containing the 'main title' and a list of the equivalent terms, so one line would look like:
SN: SN, Shane N, S North, Shane North, Shane, Credit Mgr, Credit Manager
You add and remove lines as desired. Not unlike the 'autocorrect' list you see in Word.
Then, in your searches, there would be a checkbox saying "include synonyms." If checked, a small bit of code before the existing search code looks at each term in the search and if it's in the synonym list it replaces the text simply ie:
"Credit Mgr" => "SN" OR "Shane" ... etc)
If it had %'s or other things around it those would be duplicated for each of the OR's. Then the search code runs as normal.
This is really a simple but effective implementation of the 'semantic search' ideas that have been kicking around for a while but are poorly implemented normally. This same functionality can also allow you to define 'sets' of items that sit under a concept, without the hassle of implementing complex relationships between items. For example you could define:
EmailApps: Thunderbird, Outlook, Gmail
Cars: GM, VW, VolksWagen, Ford, Jaguar.
Then in your search you could say "like EmailApps" and pick up all those items that contain a reference to Gmail etc. You could even put the headers of the sets in a dropdown using some existing functionality. You'd probably have to decide whether this could only work on the quick search or if it could work easily with the other filters and things.
What do you think everyone? If no personal examples for yourself have come to mind while reading this, I've stuffed up my description of the feature! :)
Comments