C Consiglio Nazionale delle Ricerche

(1)

C

Consiglio Nazionale delle Ricerche

Experience in implementing a Document Delivery Service

Experience in implementing a

Document Delivery Service

Francesco G ennai, M arina B uzzi,

Francesco G ennai, M arina B uzzi,Laura A bbaLaura A bba

Istituto per le Applicazioni Telem atiche

{Francesco.G ennai, M arina.Buzzi, Laura.Abba}@ iat.cnr.it

Silvana M angiaracina, M aria G razia B alestri

Biblioteca Area della R icerca Bologna {m angiaracina, balestri}@ area.ba.cnr.it

D L00 D L00

(2)

C

• The National Research Council (CNR) is a

governmental organization dedicated to the

development of scientific research and

technological progress in Italy

• 220 Institutes

(3)

C

The biblio-MIME project is a collaboration

between IAT (Institute for Telematic

Applications) and the library of the Bologna

CNR Research Area

(4)

C

Motivation

• The library of the Bologna exchanges about

5000 docs/year with other national libraries

(2000 with CNR libraries)

– The average length of a document is 10 pages

• This number of requests is managed with a

very small number of library operators, who

often change over time

(5)

C

Objectives

• to simplify the Document Delivery activity

• to improve the quality of the service, facilitating distribution of large documents

• to reduce the time of the Document Delivery process

(6)

C

Internet Document Delivery Service (IDDS)

based on an e-mail/web gateway

We describe our experience in the design

and implementation of this gateway

(7)

C

In the electronic mail we can identify two

components

- The transport protocol SMTP (Simple Mail

Transfer Protocol) used to send messages

- The mailbox access protocol such as POP (Post

Office Protocol) and IMAP (Internet Message Access Protocol) which enable users to access their own remote mailboxes

(8)

C

Sending large files can present some problems

• Mail servers can impose message size limitations

• Mailboxes access protocol time-out. In transferring

large files time-out parameters (present in the POP client) can interrupt the connection making it impossible to download the message and those following it

• The availability of remote and local resources

(such as disk space) can influence the behaviour of the IDDS

(9)

C

Idea->to combine the e-mail service (for

document transmission) with the web service

(for document downloading)

– Downloading documents using a web browser is more flexible and reliable than using an e-mail client

– The user can move between different platforms without requiring any client reconfiguration

(10)

C

MIME (Multipurpose Internet Mail Extensions,

RFCs 2045/6/7/8/9)

• MIME is an extension of the electronic mail

message format (RFC822) allowing the inclusion of multimedia data in a standard message format

• it redefines the mail body format in order to include non-textual data and structured body parts

(11)

C

MIME

Data types: text, image, audio, video, application, message and multipart.

For each data type more subtypes can be defined (image/tiff, image/gif, ...).

A multipart MIME message is a structured message composed of multiple parts containing different data types.

(12)

C

(13)

C

How does the system work?

• The library operator receives one request, locates and retrieves the document from the library and e-mails it back to the sender

• On fly the e-mail server extracts registered document types (.gif, .jpg, .pdf, .tiff) contained in MIME message parts and saves them in a web-server disk file. The extracted part is replaced with a brief service description and the link pointing to the saved part. The other message parts remain unchanged

• The resulting multipart message will continue the network trip towards its own final destination

• The receiver is able to download the documents using a user-friendly web browser

(14)

C

(15)

C

The e-mail server:

• we host e-mail services for different CNR

domains,

• we offer a MIME attachment conversion

service,

• e-mail to fax gateway service,

(16)

C

• In order to have messages processed by our

system, two conditions must be met:

– the IP address of the client has to be registered in a service authorization list;

– the e-mail destination address must end with the

fake top level domain (TLD): .save This fake

TLD domain (added by the librarian) will be automatically removed by the e-mail server. This condition permits operators to switch between regular e-mail operations or IDDS operations

(17)

C

Security considerations

• Documents stored in the web area are not protected

– listing of the dir is not allowed

– the file name is automatically generated by the system

• Documents remain in the web directories for a period specified by a system parameter. All system activities are logged in, and accessible from, protected web pages

(18)

C

The Bologna Library offers the Internet Document Delivery Service free of charge.

Libraries using the service agree to respect copyright law

(19)

C

Conclusion

Our experimentation demonstrated:

• Increased efficiency of the service, allowing the

library to satisfy all requests received on the same working day

• The service is user-friendly - no training is required

• The system offers robustness in delivering and

downloading documents, thus increasing the quality of the service

• In this specific case, the solution is cost effective

(20)

C

Document Delivery costs in 1999

0 50 100 150 200 250 janu ary febr uary ma rch _april _may _june july augus t sept em ber octo ber nove mber dece mber # of doc u m e n ts 0 100000 200000 300000 400000 500000 600000 700000 800000 900000 1000000 co st s surface mail e-mail fax Total cost in Lira

(21)

C

This proposed system could be advantageously adopted not only in the field of library science but also in any application where automatic filtering and/or elaboration of message parts is required

(22)