This page demonstrates distinctive fast autocomplete features:
- it is really fast, demo works on Celeron/300, it takes less than 1% of processor time. See sysinfo here
- it consumes low memory
- it significantly reduces regular AJAX client-server traffic,
as it sends compressed data
- supports aliases, dynamic abbreviations
- highly customizable, can retrieve different quantity of suggestions,
basing on configuration
- pattern search algorithm can be separated and integrated into
3-rd party systems
This demo contains 4 input fields, associated with 4 dictionaries:
- US Cities. This dictionary contains approx 26,000 US cities names. Suggestions in the pull-down window ordered by
city population, descedant.
- Expired domains list, for resale. This dictionary contains approx 380,000 records.
- Russian Cities. This dictionary contains approx 1500 Russian cities names.
- All Cities. This dictionary contains both US and Russian cities names, ~28,000 city names.
Fast AutoComplete is multilingval, too. The input fields below demonstrate this feature (works in Mozilla firefox only now):
- Russian Cities. This dictionary contains approx 1500 Russian cities names in Russian. Try to enter Москва (without quotes to see available suggestions in Russian).
- Mixed mode. This dictionary is a combination of the above "US cities" and "Russian cities" dictionaries. So it contains cities names both in English and in Russian.
This autocomplete subsystem initially had been designed for works in production since 2002 on
PubMed search engine in 2002. It still works
here, and serves two huge dictionaries:
Journals (~20,000 journals) and Author names (~12,000,000 personal names).
Also, this fast autocomplete subsystem serves WEB-site of Arlington Heights Memorial Library.
This demo runs on Compaq DeskPro EN: Intel Celeron/300mHz CPU, 288mb/66mHz RAM :-)
Interesting? Want to get? Just contact me!
© 2009 Oleg Khovayko