Page 66 - Visions of the Future Chemistry and Life Science

World champion chemists 55

been growing very rapidly, but it is unlikely to grow without limit. The
Cambridge Chemistry WWW server now handles about seventy thousand
requests for information each week and has been running for five years.
After two years, the growth in use appeared to be approximately exponen-
tial, and so it was possible to estimate how the load on the server would
increase. Based on just two years of data, the general shape of the follow-
ing three years of growth was predicted with surprising precision, despite
the constant addition of new material and new techniques. When the
growth of the internet levels off, the access to this server is also likely to
level off. A recent report suggests that this may happen as early as 2003,
with around fifty million computers connected together. This suggests
that accessible computer power is growing at many times the Moore’s law
prediction, but it is unlikely to continue to do so for very much longer. It
will not give the thirty or so orders of magnitude that are required in order
to solve organic synthesis by a brute force approach.
The internet allows the linking of computers which are tuned for data-
base searching (and which may access a world wide database of informa-
tion, which is not limited by the published literature but also includes
research results which are available only on the internet) with computers
which are capable of calculating chemical reactivity. It is now easy for me,
for example, to do different sorts of literature searches on computers in
Bath, Daresbury, and in Manchester, and to analyse the data using comput-
ers in Cambridge, all without leaving my office.
The next step would be to allow computers which can calculate chem-
ical properties to interact automatically with computers which can search
the chemical literature. This would enable the literature results to be
extended to the precise systems of interest for a particular synthesis. If a
new alcohol is being oxidised, then the effect of the surroundings could be
calculated, while the experimental protocol could be taken from the paper.
Thus, the literature results would guide the calculations. The calculations
would also guide the literature searching, because the calculation may
suggest a side reaction which could be checked in the literature. Literature
precedent may be a more reliable guide than calculation as to which of
several possible reactions is likely to work best.
It is only just becoming possible to use information technology to rou-
tinely search the chemical literature and to do chemical calculations
which are directly useful to synthetic chemists. Each of these fields is
likely to develop in a powerful way in its own right over the next decades.