US Department of Commerce/National Oceanic and Atmospheric Administration (NOAA), PGAS, Procurement Operations Division, 1325 East West Highway, Station 4301, Silver Spring, MD 20910-3283

70 -- AVANCED SEARCH FACILITY SOL 52-DKEA-7-90010 POC Joel L. Perlroth, Contract Specialist and Diane C. Husereau, Contracting Officer (301) 713-0829 The U. S. Department of Commerce/NOAA/Economic Affairs/ STAT-USA intends to procure software development to provide an Advanced Search facility to support an effecient means for gathering and indexing Federal Information on the Internet. The system will deploy automatic indexing of the content of sites rather than the collection of all pages into a single index. The aggregated index will also include the contents of the Government Information Locator Service compliant locators (GTLS Application Profile as specified in FIPS PUB 192) at each of the Federal agency sites. One of the key characteristics that this Facility must have is the ability to select and retrieve an original document as opposed to other versions of it or discussions about it. The retrieved result set must rank the original first in the list of returned files or in an order specified by the user. Specified fielded requests will take precedence over probabilistic determination of ranking. Components of the facility include a gatherer service, broker service, indexer, query processor, and a user interface. Gatherer: A gatherer service that extracts documents from network servers. Gathering uses a browsing metaphor that given a starting point, will follow links in html and directories in ftp collecting documents on the way. Z39.50 has a search and retrieve metaphor that given a query, will find all documents related to that query. This means that the gatherer will need to use a non-browsing technique to gather from general Z39.50 servers (e.g. a library database). Broker Service: A broker service will collect information from many gatherers, to build an index of widely distributed information. The broker service component of the search engine will use the indexer service to create and manage the indexes. Indexer: An indexer service will parse gathered documents and identify index terms representing the content of documents. The system must support a variety of indexing techniques including: simple word-based indexing, indexing based on part-of-speech tagging and phrase identification, indexing by domain-dependent features such as company names, dates, locations, etc., and fielded tag-value pairs. The indexer will create GTLS compliant locator records for any documents that do not currently have GTLS compliant locator records. Query Processor: The system must integrate natural language, Boolean, and proximity queries, including field-based retrieval. The system must represent and use spatial information in indexes and queries. A part-of-speech tagger is to be used to identify candidate search phrases. User Interface: Must allow alternative forms-based query specifications: natural language and fielded-boolean. Both of these alternatives must exploit all features of the search engine, including relevance-ranking, feedback and best-passage highlighting. The locator records shall be person-editable so that additional information can be added to the locator records in order to manually-weight the ranking process by system administrators. This will allow the versioning and chain-of-authority features of GILS to be used in ranking search results. An ''original'' document must appear first in a ranked list if it can be identified as such. The original and all of its versions must appear before documents that reference it. The ranking order must allow for newest-version first ranking of ''original'' documents. Implementation: Prototype software must be installed and tested at two sites on the Internet. Two sites are needed to test and demonstrate the distributed indexing and search capabilities of the prototype system. The government will provide access to two or more Z39.50 V2-compatible GTLS compliant servers in order to test the ability of the system to search, index, and retrieve documents referenced in a GTLS record on the network. Support: The Advanced Search Facility will require server and telecommunications capabilities necessary to support the Facility's search engine and a highly-active Internet World-Wide-Web site. The period of performance is for one (1) base year for software development with two (2) one year options for software support. This procurement is a 100% small business set-aside, (See Note No.1). Also, this procurement is not being conducted in accordance with FAR Part 12, Acquisition of Commercial Items, (See note No. 26). The SIC Code 7371. Copies of the solicitation will only be provided in response to written requests. Written requests may be faxed to (301) 713-0806. Telephone requests will not be honored. (0215)

Loren Data Corp. http://www.ld.com (SYN# 0297 19960805\70-0003.SOL)