Showing posts with label ostracon. Show all posts
Showing posts with label ostracon. Show all posts

11 May 2006

The Digital Sum of Human Knowledge

Most of us think of open access as a great way of reading the latest research online, so there is an implicit assumption that open access is only about the cutting edge. This also flows from the fact that most open access journals are recent launches, and those that aren't usually only provide content for volumes released after a certain (recent) date, for practical reasons of digital file availability, if nothing else.

This makes the joint Wellcome Trust and National Libary of Medicine project to place 200 years of biomedical journals online by scanning them a major expansion not just to the open access programme, but to the whole concept of open access.

It also hints at what the end-goal of open access must be: the online availability of every journal, magazine, newspaper, pamphlet, book, manuscript, tablet, inscription, statue, seal and ostracon that has survived the ravages of history - the digital sum of all written human knowledge.

07 January 2006

Code is Law, Code is Politics

As Lawrence Lessig famously noted, Code is Law. Which means that Code is Politics, too, since laws are drawn up by politicians. But the intimate relationship between code and politics is becoming manifest in a rather different context (pity about the yellow on black text).

The issue here is about the software used in voting machines. Since, one day, all voting will be carried with such machines (unless we decide to go back to using ostraca), now is the time to consider why free access to the code that runs them is indispensable for political transparency.

It comes down to this: if you are dealing with a black box, you can have absolutely no faith in the results it produces. It might just make them up or - worse - change them subtly, or perhaps be pre-programmed to crash if a particular party gets too many votes, requiring a complete re-run, with knock-on effects on voting patterns.

If you have the source code you can run it and examine what it does with various voting inputs, and check that it has no nefarious sub-routines. However, even this is not enough for full confidence in the voting machine: paper audits are also indispensable for checking on the consistency of the outputs, and allowing for the ultimate fall-back - counting by hand.

Still, this is a clear instance of where, in a literal rather than metaphorical sense, closed source jeopardises the very basis of democracy. Looks like RMS was right.