The Definition Server



The Definition Server




Related Areas
Wireless
The Definition Server


For most of the information required on technical topics, the learners have to run around and search through hoards of books. This consumes a lot of time and effort and the learners often feel the need for an autonomous system that can handle such queries. Till date NIIT has pioneered many strategies to improve learning effectiveness. As another step in this direction, CRCS has built the Definition Server; a Web Bot based system to help the learners get an on-line answer to the queries related to technical topics.

The Definition Server is an autonomous system that can handle technical keyword related queries. The system receives a query for a definition in a POP3 mailbox from a standard e-mail source or as an SMS mail from a cell phone. A mail handling system is used to scan the mailbox periodically. On receiving a mail with correct query syntax, the response is picked up from the definitions database. If the definition is not found in the database, the word is sent to the queue of a Web Bot. The Web Bot looks up for the definition on the Internet and on finding it populates the definition database. As soon as any queried definition is found or put into the database, it is sent back to the user as an e-mail/SMS message.

As an extension to the mail handling system, a PSTN based Interactive Voice Response System (IVR) has also been added. The definitions stored in the database, are converted to audio files using a Text to Speech engine based module in a batch manner. This allows the user to ask for the response either in a text form or as a voice call playing the audio file of the definition. It is even possible to initiate a query from one source and get the response redirected on any other ordinary/SMS mail, PSTN phone or wireless phone.

The whole Definition Server system has been distributed across several modules, which may be present on a single machine or may be distributed across multiple machines. The distribution on multiple machines gives some degree of tolerance to the system in situations of hardware failure. This has made it scalable to handle higher load requirements. The various distinct and tangible modules are:

a) Reception Server - Receives/parses mails from POP3 account to populate query database.

b) Email Responder - Sends definition text to an e-mail ID using SMTP protocol.

c) IVR Responder - Sends voice calls over PSTN line.

d) Web Bot Server - Searches/updates pending definitions from database.

e) Text to Speech Converter - Generates audio files for all definitions

f) Administration - Allows a manual operator to monitor and adjust information in the database, enter a definition manually, invoke audio generation

Most of the system has been programmed in Visual C++ and Visual Basic. Various technologies and protocols that have been used are TCP/IP, HTTP, POP3, SMTP, CTI, TTS, and ODBC. Few aspects that are still being looked into are:
  • Automatic conversion of generated audio files into 8 bit 8khz mono
  • Affiliation of Internet resources to the Web Bot to increase its search capability
  • Integration of manual recording of audio in the Administration module

This project reflects the efforts of CRCS in the area of Internet technology.


For more details contact: Swami Sachidanand