reviewers. [10] These contributions are most often done on a volunteer basis though it could be argued that their respective institutions pick up the tab for the time they spend on the journal or reviewing submissions. Yet even if we were to factor in the cost of the volunteer editorial and review functions, the benefits of handling texts electronically would still reduce the cost in comparison to that associated with paper publication. As for electronic storage and transmission costs, these are now quite trivial. In 1994, Paul Ginsparg (1994) noted that cost for gigabyte of storage was under $700. This meant that the 25,000 physics papers published each year could be stored for about 3 cents apiece. Since that time the cost for a gigabyte of storage has plummeted to about $300 a gigabyte thus further trivializing the cost of storage. However even in 1994, Odlyzko could conclude that the cost to store all current mathematical publications would be less than the subscription cost for one paper based journal! As to the cost of internet connects, these are generally shared among all members of an organization. Odlyzko (1994) noted that even with the recent withdrawal of NSF support for the Internet infrastructure and the move to commercialization, academic storage and transmission should remain trivial because network transmission will have to remain cheap enough for commercial applications (pictures, movies, etc.). He concludes by noting that the cost of fast internet connect will remain less expensive than the cost of a good collection of paper journals for only 1 discipline.Estimates as to the cost savings of publishing material electronically range from a low of 25% of paper based publication costs to a high of 75% (Garson, Ginsparg and Harnad, 1994). The variation in estimates seems to be discipline specific. The cost of publishing humanities journals where typesetting requirements are minimal is lower than say publishing chemical jou...
