|
|
LinkBaton Background |
LinkBaton is based on a simple and old idea, that things should be called forth by name, not by retrieval method. We'll find a good verse in Genesis to cite.
Very early in the development of the World-Wide Web, the notion of accessing things by Uniform Resource Names (URN's) was thought to be a good thing. Although people thought about how an URN system might work (see RFC 2169) and URN capable systems have been deployed (see the handle system) for a variety of reasons the use of URN's at the protocol level has never caught on.
Perhaps the difficulty with accessing information by name is that names have lives outside of the internet. Witness the difficulty with domain names and trademarks. Nonetheless there are names (identifiers) that successfully cross the barriers between the real world and cyperspace, such as International Standard Book Numbers (ISBN)'s. More problematic have been useful resources such as scientific journal articles. Should they just be given ISBN-style numbers (such as a Digital Object Identifier), or some sort of code computed from bibliographic data (such as SICI)?
Recent work in presenting multiple sources of information to users has been a direct inspiration for LinkBaton. Herbert Van De Sompel's SFX among other innovations, showed the power of parameterizing link destinations in user cookies. The National Library of Medicine's PubMed Linkout showed the power of centralization in links for named resources.
Some of the most innovative uses of cookies and linking are to be found in the internet advertising and e-porn industries. Careful study of the techniques used by DoubleClick, BeFree and LinkShare are highly recommended to the student of Linking.
LinkBaton combines ideas from many sources and adds some innovations of its own. We hope it will be useful.