New Meta data based Linux desktop search system, which overcame the limitation of current Linux desktop search capabilities, was developed. Thus likely to fuel more active developments in open source based semantic desktop search technology too.

On 21st, South Korean R&D firm ETRI announced that they have developed new Linux desktop search system, 'antbear'. The antbear can analyze and run advanced text searches, producing more accurate results by precisely analyzing the search words. The system is an enhanced Linux desktop search, maximizing utilization of meta data (ID3V1, ID3V2, EXIF, IPTC).

Advanced word recognition technology gives more reliable search results
Linux desktop search, thus far, had some limitations due to its lacking features compared to Windows environment that has various features. Although 'Beagle' has been available for supporting Korean and other Asian languages (Chinese, Japanese), its 'bigram' search method in many situations fails to recognize true meanings of Asian characters for the search. For example, it breaks down and indexes the search word(s) into two separate characters, which can generate unrelated results. The reliability of the search result gets even worse when its search is based on larger number of files.

But antbear delivers solution to where Beagle left off. Using its advance word analyzing feature, results irrelevant to user’s query are omitted in the search, giving more accurate and reliable results.

The antbear supports document searches in both Korean and English and even analyze other Asian languages like Japanese, Chinese with simple plug-in. The system can run on any Linux kernel version 2.6.13 or above.

Expanding search capabilities using various metadata
Another heroic of the antbear is its search capabilities in vast areas with Meta data ID3V1, ID3V2, EXIF and IPTC. Unlike conventional search systems with limited areas (file size, location, modified date) of Meta data search using file name. When searching picture files, the antbear also analyzes in many different aspects of the image: number of persons, color or black/white, indoor or outdoor. Using this Meta data, user can run more specific query for the search (scenic pictures or self portrait, etc.)

ETRI voice information center/ Knowledge mining research team leader Myung-Gil Jang who's in charge of the antbear project said, "Globally, current trend of search systems are framed using semantic desktop search (for sharing PC base personal knowledge using RDF schema: Resource Description Framework). ETRI's goal is in developing semantic desktop search system, which could come up as the most important technology in next generation of PC environment."

"Development of the antbear brings Linux to be more user-friendly, which enables new areas of opportunities for the future software developments in Linux desktop industry." Youngmin Yoo, director of KIPA (Korea IT Industry Promotion Agency) pointed out.

Meantime, ETRI earlier in May had developed Linux personal media management system that includes desktop search.