I agree that some human review is necessary, but it can take much more
distributed forms than a strict hierarchical model. I also recall somebody
telling me that in some crowdsourcing projects, when two people
independently agree on something, they are so often right that it is not
worth checking further. A high percentage of the more obvious defects in
the TCP texts fall in that category. I think that in the Old Weather
project contributors got performance based privileges and could move from
cadet to admiral.
Professor emeritus of English and Classics
On 2/22/15 1:58 PM, "Sebastian Rahtz" <[log in to unmask]> wrote:
>> On 22 Feb 2015, at 19:17, Fabio Ciotti <[log in to unmask]> wrote:
>> Speaking of problems, such
>> a system should have a complex system of user right control and
>> proposed editing supervision.
>It is mildly depressing that peoplešs thoughts immediately turn to control
>and permissions and supervision and what have you. Has 20 years of the
>movement taught us nothing? that a) in general you can trust people,
>that b) systems settle down and reach a stable level, and that c)
>provenance is more important than restricting change.
>Chief Data Architect
>University of Oxford IT Services
>13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431