| UNIVERSITY OF PENNSYLVANIA - AFRICAN STUDIES CENTER |
*My subject line and title are borrowed from the title of a paper by Larry Press submitted at the INET'93 conference.
Although network access is expanding rapidly in developing countries, one of the largest problems that remains is obtaining reliable and inexpensive access to the wealth of information and online resources available through the Internet A wide range of services could be made available to users on low- bandwidth networks (and networks with only periodic access to the Internet, as is common in developing countries), if online databases, archives and other services used some form of email-queriable response system. Email-queriable response systems could be designed for many internet services including access to library catalogs, indexes and other databases, gopher searches, ftp, and so forth. Universities and university departments could make archives of working papers and other current research available to email queries.
The list of potential resources that could be made available can be quite large. Using a digital signature system based on Certificate Authorities (CAs as outlined in RFCs 1421-24), more restricted resources such as CD-ROM databases could be made remotely accessible while limiting access to a predefined community (as I understand it, this is one of the main limiting clauses in CD-ROM purchase and subscription clauses). A digital signature system could also be used to provide secure remote access to home mail boxes via email queries. This method could also be used to support paid subscription to electronic versions of journals. If the volume got high enough, journals could actually offer discount subscription rates to people who subscribe in this fashion due to savings in materials (paper & printing costs) and mailing--and users in developing countries in particular could get their issues much more quickly than by postal mail.
Developing Standardized Email-Queriable Resources
This tool would be even more useful if a standard set of formats and procedures could be agreed upon (perhaps by an Internet RFC), so that it could be widely implemented in a common format. This would also allow the development of easily shared server and user-interface software (which could be developed for multiple platforms, such as IBM-compatible PCs, Macintoshes, Unix workstations, etc.).
The work and funding necessary to develop these kinds of resources would not be great. Most can be done with existing utilities and tested software, repackaged in a way to make it easy to use. I know of or have heard of some existing systems already, such as the Comserve database (Comserve@Rpitsvm.BITNET) at RPI in Troy, New York (which is a large, email-queriable archive of documents and other resources for the International Association for Mass Communications Research (IAMRC)), and assorted ftp archives that can respond to email queries. Chasque in Uruguay, one of the APC (Association for Progressive Communications) networks has put a lot of the documents from the UNCED conference in Rio online and accessible via email query. Also, SateLife out of Boston disseminates a wide range of medical information to developing countries, and has an agreement with the publishers of the New England Journal of Medicine to allow rapid dissemination of current articles.
Arguments for developing Email-Queriable Internet Resources
I think that the main focus needs to be on developing easy to use resources that can receive and process requests via email, and return responses and output in ASCII output--whether as ASCII text files, or as UU/XXencoded files (with or without compression using some popular form such as ZIP, and with the option of splitting the output into files of a specific size to work around some networks' constraints). The software developed to do this could be designed in a modular fashion, so that new advances, as they are developed (such as MIME), could be easily added as options. The primary design criteria should be how to deliver the widest range of resources quickly and easily to the widest possible audience, regardless of where they are, what kind of network access they have, etc.
Many that I have talked to about these kinds of ideas say that it is a good idea, but that networks and transport resources are developing so quickly that any effort put into developing this kind of delivery system and applying it to a range of Internet resources (like gopher and ftp) would be obsolete before it even really got off the ground. To someone accustomed to network access in the United States and Europe, this *might* seem to be the case. However, there is a very strong case to be made to the contrary--for many reasons:
1) Even when Internet access has been developed in other parts of the world, access to national academic networks is often still limited, so many potential users have to turn to other kinds of networks (Fido, UUCP, etc) to get access now--and probably will have to for some time to come. This kind of service would open up a wide range of resources regardless of what kind of network one is on. 2) Access--Internet and otherwise--is still a very expensive activity, and on-line time can be a significant portion of costs, particularly if access is over commercial packet switching networks (email queries are cheaper in this context). 3) Email queries can be spooled and processed when network traffic is quieter. This has an advantage not only for networks which connect to the Internet on a periodic basis, but also for the institutions that provide this kind of access. Far more people per day can make requests to an ftp archive this way, than can through a direct ftp connection, and existing resources can be used more efficiently. This can actually free up more bandwidth for the 'sexier,' higher bandwidth resources. 4) Building a baseline system that has, as its primary directive, accessibility to the widest possible audience from the ground up is not antithetical to new technologies and techniques. By raising such issues and designing systems with them in mind, new technologies can be easily incorporated as new options when they become available. 5) This kind of resource can also be used now to facilitate the transfer of information from 'South' to 'North,' or 'South' to 'South,' as it can provide a simple means for institutions on networks outside the Internet or having asynchronous (periodic) connections to the Internet to place information online and make it widely accessible. The means could be developed quickly, using existing technology and software that is well tested in the public domain. If the resulting server software is also placed in the public domain, it could be readily adopted in developing countries. 6) By far the largest group of network users are most accustomed to using electronic mail as their main medium for communication and information retrieval. Email-queriable resources, if well designed, could open up a far wider range of services and information to users in an already familiar environment (their local mail program) without having to learn a large number of specialized programs. Thus they can be given ready access for a minimal investment in training.
Developing both Server and User-Interface Software
As I mentioned above, this approach has implications not only in how resources are made available to potential users, but in the design of user-interface software that will facilitate access to these kinds of resources. User interface software can be designed to produce routine (or complex) queries that are in some canonical ASCII form, which the user can then paste into an email message and send on. Or if, for example, a user has a network account that does not allow uploading and downloading, the user-interface software could allow them to print out the generated query on a desktop computer, and then hand type the message into an email message on their network account. Naturally the standard format for queries should be simple enough for users to type their own commands when familiar with the system, but such a feature would help newer and less sure users.
A software package like this could also include integrated utilities (like compression + UU/XXencoding combined in an Interface that is easy to understand and use), and could be developed by a consortium of universities for use on a wide range of computers, and distributed as freeware. Commercial versions could add more utilities. Specialized versions could generate time-stamped digital signatures (valid for only a specified time period, for example), which could then be appended into an email message, allowing access to more limited resources. Similarly, the server software developed to run email-queriable systems like this could be developed for multiple platforms (Unix, VMS, CMS, DOS, Macintosh, etc.) and placed in the public domain, so implementation of this kind of system can be wide-spread and as uniform (standard) as possible. Note: both software packages (user & server) should support multiple languages. Comserve is exemplary in this respect.
Those are just my ideas on the subject, anyway. I would be very interested in entering into a discussion of these issues with others, and to learn about projects that might already be trying to implement these kinds of resources. I would also be very interested in hearing more from potential users, particularly those on networks that are not directly connected to the Internet, or whose network connection travels over an asynchronous link (such as periodic dial-up or satellite). In particular, I would like to participate in a discussion of:
1) What kinds of resources would be useful if accessible in this form. 2) For what community of users (Academic, NGOs, individuals, libraries and other information providers, etc.). 3) What kinds of design features could make it more useful to users in developing countries (such as the capacity to send compressed and uuencoded output, the ability to split files into user specified lengths for use on networks with strict messages size limits, MIME support, etc.) 4) How standards for this kind of delivery system could be developed (or what work has been done on it already). 5) How consortia could be developed to design and implement systems like this, and how they could build on existing work to make the job easier. 6) How to facilitate development and distribution of the software.
Please feel free to repost this message, and make comments or suggestions.
Ned Bade, eb15@cornell.edu