[MKSearch-dev] MKSearch as a Desktop Application

[Phil Shaw]

On 6 Aug 2006 at 8:19, Perl Dog wrote:

> Hello, I would like to use MKSearch as a desktop application; to be used by
> many individuals on their own computers, not installed on a central server.
> 
> How practical would this be?  I could have a centralized mySQL database if
> needed, but I would want the BULK of the bandwidth to use used at the
> desktop, with minimal bandwidth requirements for the server.
> 
> Can this be done?


There are several options for making a desktop search tool with 
MKSearch. Only one is currently implemented, but it can be deployed 
in different ways.

The system is currently implemented as a Web application with Web 
browser clients. The servlet container may be on a remote host, 
separate from the client, or it could be on the same machine. 
Likewise, the database host could be local to the servlet container 
or remote, and a number of installations could access the same 
database.

In an intranet type deployment, any number of users could connect to 
a single servlet container in a similar way to an Internet 
deployment. The single central servlet container would use a single 
database repository.

For a single user, the servlet container could be deployed on a 
workstation and accessed using a Web browser on the same machine. 
Another option would be for servlet containers on numerous desktops 
to access the same central (intranet) database.

The other main option, which is not currently implemented, would be 
to make a graphical user interface client for the system. This might 
connect directly to the repository database.

Command line options are also possible. There is a basic 
com.mkdoc.sesame.RepositoryQuery class for testing SeRQL queries that 
uses a text file for query input against a file-based Sesame 
repository.

Before the initial beta phase of development ended, I started work on 
a com.mkdoc.io.FileSpider to test the principle of indexing content 
on local or intranet file systems. This was mainly for test indexing 
files locally, but could be expanded for use in earnest.

Best regards,

Phil
--
MKSearch (beta)

http://www.mksearch.mkdoc.org/

Free, open source metadata search engine with RDF storage and query.