From Sean Barker
In speculating that we might lose vital knowledge if civilisation were to collapse, your authors missed the very real threat that it might happen in the near future (30 January, p 36).
Most software is updated every few months, and is frequently completely rewritten, which can result in errors in reading the old programs. This can be a real problem for the engineering systems in planes and ships, for example, where accurate information is needed to maintain and repair them decades after they are built.
And what about the computer models that prove systems are safe? In years to come, the models showing that the Large Hadron Collider is safe may no longer be trustworthy after several generations of upgrades to the software.
Printing out the relevant information or stockpiling old computers will not help, since people in the future will not only need the data, they will have to understand it. Aside from the software side of the problem, there is an even more difficult question: do we know how to train people to understand the old models?
Advertisement
My own work in the development of a standard for the long-term archiving and retrieval of aircraft product data suggests that these problems are not as rare as the article suggests.
Bristol, UK
