Review
------

'As We May Think' by Vannevar Bush

In the article 'As We May Think', the author, Vannevar Bush, holds up an
incentive for scientists in the post World War II era. In order to handle the
sheer volume of information becoming available to research scientists, Bush
argues for the need for new methods of data acquisition, storage, and analysis.

For the purpose of data acquisition he proposes the use of a head mounted camera
and dry photography. He proposes extending the technologies that had already
been realized: facsimile and television. Bush addresses the issue of storage by
counting on microfilm technology motivated by the fact that it provided an
inexpensive solution. He suggests dictation technology as a means to extend the
written knowledge and an extreme approach of devising a natural language suited
particularly for analysis and parsing by the machine.

Bush does a great job of predicting the advanced capabilities of computers (of
the future) in the fields of advanced mathematics, logic problem solving, etc.
He also recognizes the essential need for digital indexing of documents for the
automatic retrieval of the same from mass storage.

He introduces memex, a hypothetical, desk-size information workstation, which
contains a massive amount of microfilm that stores documents. In this system,
documents could be recalled by document code and page number and displayed on a
screen. His key observation was that of associative annotation and selection:
arbitrary pages could be linked together to form a trail of documents. This is
probably the first description of hypertext, which forms the basis of the World
Wide Web today.

A key aspect to note is that Bush does not consider/discuss digital content
storage. Since Bush's notion of storages was not digital, he has had to rely on manual
indexing of documents. Also, his design lacks the ability to search for
'keywords' in the knowledge system. Each individual using the memex system
maintains a local copy of the knowledge base (library data, etc) and is
responsible for constructing trails. This is clearly undesirable as the same
trail might have been to be created several times over. He presents a poor
model of knowledge dissemination: users share trails only by hand. But with the
help of networking, publish/subscribe systems, and search engines, today's
users can share and access information at lightning speeds presenting a much
better model of knowledge dissemination.

In conclusion, Bush does a good job of proposing novel solutions to create and
organize data and predicting capabilities of technological devices of the
future.