EPP Contents

Disclaimer

This software has been developed by Andreas Eggenberger and Rolf Stadler at the CTR, Columbia University. It is to be used for non-commercial purposes only; it may not be used for or incorporated in commercial products or services without the prior written permission of the CTR.
Software and information furnished on this page is believed to be accurate and reliable. It is distributed as is, with absolutely no warranty of any kind, either expressed or implied, including, but not limited to, the implied warranties of merchantability of fitness for a particular purpose. The entire risk as to the quality and performance of the programs is with you. Should the programs prove defective or cause any damage, the CTR assumes no responsibility to that cost or damage. The CTR also assumes no responsibility for any infringement of patents or other rights of third parties which may result from its use. No license is granted by implication or otherwise under any intellectual property rights. Software and Information is subject to change without notice.

Center for Telecommunications Research, Columbia University, New York 1996

EPP Design Concepts

Phases for paper processing

Paper submission by authors
Paper assignment to PC members
Review process by PC members and reviewers
Evaluation of the reviews

Data acquisition by servers

Paper information from managing author

via email/fax
ftp/mail

Responsible TPC members from PC co-chairs

via WWW

Reviews from PC members/reviewers

via WWW/email/fax

Data entities

Submission phase

Paper information
Title, author(s), abstract
Managing author's address
Ps-file of paper

Paper assignment phase

Information on which TPC member is reviewing which papers

4 TPC members per paper

Review process phase

Reviews

Each review contains

TPC member's and reviewer's identification
Overall evaluation
Topic classification
Comments to Program Committee
Assessment
Comments to author(s)

Evaluation phase

Paper ranking
Statistics

Data formats

Postscript-documents (papers)
HTML-files (different views)
ASCII-text files (for exchange between different systems)
MS Excel-spread sheets (ranking, statistics)

Web views and accessibility

PC co-chair's view and administrative view

Paper database with info on managing authors
Responsible TPC member
Reviews
Statistics
Review summary

TPC member's view

Paper database
Some of the statistics
Review summary

Public view

Submission guide
Submission form
Review form (for reviewers who are not TPC members)

Software components

Common Gateway Interfaces (CGIs)

Accept processes information received on the paper submission form and generates html files for the abstract, the managing author's address and the overview of all the submitted papers.
AssignReviewer processes information received on the reviewer assignment form and generates an html file per paper which contains the TPC members responsible for this paper.
AcceptReview processes information received on the review form and generates an html file.

Filters and other tools

CreateAllEntries creates a table with one entry per review containing

PaperId
Title
Author(s)
Overall evaluation
Suitability
Quality
Originality
Presentation
Readability
Suitability for poster session

This table is imported as text file into MS Excel and used to compute the ranking.
Command syntax: CreateAllEntries <input file> <output file name>
The input file contains the file names of the reviews that are to be processed.

GenerateRemovedInfoReviews creates reviews without the reviewer's and TPC member's identification.
Command syntax: GenerateRemovedInfoReviews <input file>
The input file contains the file names of all the reviews that are to be processed.
ExtractAllSubjectAreas creates a table (relation) containing

PaperId
Topic

This table is imported as text file into MS Excel and used to generate statistics.
Command syntax: ExtractAllSubjectAreas <input file> <output file name>
The input file contains the filenames of all the reviews that are to be processed.

Other tools needed to convert file names, automatically generate html-code, process log files, etc.

Environment

OS/Compiler

The CGIs are written in C, the filters and the other tools are written in C++, using the developer's environment Sniff++. The CGIs run on SUN OS and the filters run on IRIX 5.3. As compiler we used gcc on SUN and the SGI CC. The programs should be easily portable to any other UNIX system.

Web server

We are using the NCSA web server. This has an influence on how the password protection for the web pages works. The web pages are HTML2-conform and contain frame sets.
The web server is running on a SUN workstation under SUN OS.

FTP server

The ftp server runs on an IBM RS-6000 under AIX.

Installing the system

In order to set up the system, all the links (URLs) in the web pages have to be updated. For instance, instead of "http://www.ctr.columbia.edu/.." the new URL has to be inserted. The web server has to be configured in a way that cgis can be used.
The ftp server must be configured such that people can upload files but cannot read the directory or download files.

Disk space

The EPP software, CGIs and tools encompass less than 1000 lines of code (source and binaries together are less than 1 MB). Most of the space is needed for the postscript-files (65 MB for the papers in case of IM'97). The total size is about 85 MB for IM'97.

Some experiences with using EPP for IM'97

We list some of the difficulties we had.

Submission phase

The cover page was mostly sent by e-mail, not via web form.
Very few authors had

connection problems due to wrong configuration on the sender side;
problems due to overload of the ftp server during the last day before the deadline (around 6pm).

Review form

For long comments, web forms do not seem to be convenient. (Web browsers are not good editors.)
Some reviewers submitted the same form several times, with slightly different reviews.
Some reviewers provided the wrong data, e.g. wrong paper id, title, TPC member's name.