The Contribution Accountant (CA) is an open-source Java application that runs on a Linux machine running an Apache server with a MySQL database backend. This may or may not reside on the same server as the Archive Assistant (AA). This application is currently only in the planning stages.
We foresee an inconspicuous and semi-transparent accounting system that will keep track of the contributions made by each transcriber in order to attribute appropriate credit to the authors that submit new or improved transcriptions. Such contributions could be rewarded by a sort of “credit” that can be used to acquire advanced services from our system as discussed above. Transcribers would also be rewarded through an automatic academic citation system that recognizes an author’s contribution as bona-fide intellectual property which must be properly referenced whenever an author’s transcription is used in a project or cited in a paper. The credit system is predicated on the registration of each user and on the system’s ability to unequivocally recognize legitimate users through passwords. After registering, users will receive a virtual “bank account” where their credits will be posted by the CA. The essence of the CA accounting system is based on before-and-after comparisons between what the AA sent out to the user and what comes back from the user for each manuscript page. A differential engine will quantify the number of changes made to the original transcription as well as to the word boxes in the manuscript image. The net change will be the basis for the credit awarded to the user. However, to avoid spurious submissions from users bent on “gaming” the system, the initial credit would be awarded on a provisional basis as a percentage of the total earned by the transaction. The “discounted” rate of the temporary credit will be based on the credit-worthiness of the user, which depends on his or her credibility, which is accrued over time, by indirect peer review.
Peers will in essence show peers their approval by implicitly confirming the quality of submitted changes, simply by not putting forward additional corrections to what the original contributor(s) had provided. A sort of tacit consent will thus enhance the credibility of users whose submissions pass muster with subsequent users. In addition to this process of “dynamic accreditation”, some archives may choose to start from an official list of fully accredited professionals to expedite the “reputation management” component of the system, using a form of “a priori accreditation”.
Our aim is to arrive at a universal registration system that will maintain the distributed and emergent nature of the overall system without compromising the quality of the transcriptions that bubble up to the top of the heap after successive refinements by a variety of contributors.
Users could accumulate “transcription credits” and “word box credits” (possibly on different “pay scales”) when they make contributions to the system. Credits could in turn be usable to pay for such services as on-demand scanning of manuscripts, handwriting recognition assistance, remote storage of project files on the archive servers, automatic transcription processing, advanced searching, and others.