| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle ConText Option
On Thu, 22 May 1997 08:47:54 -0400, roy_at_popmail.med.nyu.edu (Roy Smith) wrote:
>Anybody using ConText? Is it really as good as the sales literature makes
>it sound?
>
>Also, what are the differences between 1.1 and 2.0? From the material I
>can find describing the capabilities of both, I don't really see much (if
>any) difference in the two.
>
Here are the new features in 2.0
Indexing
URL data store
The text column can now contain URLs. All Web pages and files accessible
through the HTTP protocol from the machine running ConText Option, whether
these pages are on a corporate intranet or the World Wide Web, can now be
indexed. The URL data store supports the "file" protocol as well, to enable
local files to be referenced in the same column as URLs.
Composite primary key
ConText Option now supports text tables with a composite (i.e.,
multi-column) primary key.
Extended support for document formats
ConText Option now supports an extended set of document formats, including
HTML, PDF, Word 7, Excel 7, PowerPoint 7, and WordPerfect7
Number of stop words has increased from 255 to 4095
External filters
ConText Option now provides a mechanism by which application developers can
set up external filters that ConText Option will use to index specific
document formats.
Linguistics
Theme Indexes
Document themes can now be indexed and queried using the same paradigm as
document text. This new approach enables ConText Option to perform a more
sophisticated comparison between a theme query and the theme index, leading
to significantly better query results.
Service queue priorities
Administrators can now indicate the order in which to process linguistics
tasks.
Improved gist quality
Gists have been improved by creating more sophisticated paragraph selection
algorithms, and by the use of filtering to remove text artifacts common in
Web pages and electronic mail.
Improved performance
A typical customer can expect linguistics indexing performance to be twice
as fast as in ConText Option 1.1.
Thesaurus
ISO-standard thesaurus framework
The ConText Option thesaurus framework will comply with ISO-2788 . Valid
query terms will include the following: multi-word phrases, compound
phrases, and disambiguated homographs. Valid term relationships will
include the following: non-preferred synonymity, preferred synonymity,
broader terms, narrower terms, top terms, related terms, and partative and
generic relationships.
Querying
Improved mixed query performance (for two-step queries) A "mixed query" is one that contains both structured conditions (i.e., LAST_NAME='SMITH') and text conditions. ConText Option now allows mixed queries to be executed completely within the ConText Option PL/SQL "CONTAINS" procedure, leading to improved performance.
Maximum length of query string has increased from 2K to 4K
Support for stored queries
The results of a particular query can now be stored for access by another
query. Stored queries can also be nested within other stored queries. This
feature enables sophisticated classification and grouping of documents.
Distributed queries support
A ConText Option query can now use a database link to access a text index
located in a remote database.
NOT operator
The NOT operator allows further refinement of a query (i.e., find all
documents that contain "bank" but NOT "river"). It can not be used
standalone (i.e., it is invalid to simply search for all documents that do
NOT contain "river").
Count query hits
A developer can now use the COUNT function to request the number of query
hits. This function will give much faster response than running the entire
query and doing a manual count. There is also an approximate count
available, for even faster response time.
Incremental results retrieval using the FIRST/NEXT operator Rather than requesting the complete results of a query, developers can use the FIRST/NEXT operator to request a specific group of query hits (e.g., hits 1-10 only). This functionality allows developers more control over hitlists, and allows them to return control to the user more quickly.
In-memory queries
In-memory queries are text queries in PL/SQL that do not write their
results to tables. They return the first set of results more quickly than
table-based queries.
Parallel query execution
ConText Option can use multiple query servers to process the same query,
leading to faster performance.
Viewing
16-bit viewer for Windows 3.1, Windows NT and Windows 95 This viewer is a standalone 16-bit application designed for Windows 3.1. It will also run on Windows NT and Windows 95. It includes WYSIWYG and ASCII viewing, and will be able to leverage external filters.
32-bit OCX viewer for Windows NT & Windows 95 This viewer is accessible only through an OCX-enabled development tool, such as Visual Basic. It provides highlighting and viewing capability for WordPerfect 6, Word 6, ASCII, and HTML.
Administration
Load Server
The Load Server is a new personality of the ConText Option Server process.
It continually scans a designated directory for new files, automatically
loading any that it finds into a designated text table.
Indexing statistics
Indexing statistics are now stored in a database table.
Timestamps for server log file
Ability to resume failed index creation or optimization If an indexing job fails, ConText Option will often be able to complete the job from the point of failure, rather than starting over from the beginning.
>--
>Roy Smith <roy_at_popmail.med.nyu.edu>
>New York University School of Medicine
>550 First Avenue, New York, NY 10016
>
>Copyright 1997 Roy Smith
>For-profit redistribution prohibited
>
Thomas Kyte
tkyte_at_us.oracle.com
Oracle Government
Bethesda MD
http://govt.us.oracle.com/ -- downloadable utilities
 |
 |