frabcus
commented
12 years ago Yesterday I met with some of the people in the US Federal Government making their new FOI portal. It will, in theory, end up open source and covering large amounts of the US Government. However, while it will have a comprehensive backend cost-saving workflow, the frontend will be overburdening. e.g. Asking for postal address.
So it made me, at least in the US, even more confident that an Alaveteli here needs to be very campaigning. It should insist on 1) not paying fees, 2) emailed responses. And endlessly shame agencies that cause trouble and make things complicated by not playing ball.
Which countries seriously thinking of doing an install need this scanning version? Do they really? Surely their Governments are moving to electronic systems anyway, and a campaigning email-only Alavateli will encourage them.
My hunch would be that for the same work effort of doing the code and ongoing maintenance of this scanning... You could get a lot of PR saying how ubelievable it is in 2012 that a Government can't reply by email, when they're writing all the documents on a computer anyway.
sebbacon
garethrees
In many countries, an email-only workflow will only work as a campaigning platform.
The only sane way I can think of for supporting this is for an organisation to do the printing and scanning on behalf of individuals. It would work like this:
1) User visits Alaveteli, finds authority, writes request 2) Request is sent via email (where available) to give authority a chance to respond by email first (offering them a way to save money!) 3) After some period of time, an admin from the organisation running Alaveteli prints out the request, along with a reference number that is designed to be easily OCRd 4) Letter is put in the post, admin marks request as sent 5) Once a week, admin bulk scans any responses which are automatically uploaded to Alaveteli 6) Admin then logs into management interface. Successfully OCRd responses are already attached to the relevant request. Unsuccessful ones are put in the holding pen.
I propose that the print-and-OCR part of the application be implemented as a black box that simply consumes and emits emails, so that Alaveteli can be used as the front end without modification (this is the approach being explored by Acceso Inteligente in Chile for using Alaveteli as the user-facing front end to their form-posting-page-scraping system).
We will need to consider integration with bulk scanners (presumably quite simple, but I don't know anything about them), and OCR technology. Derek from TI Georgia has experience with Tesseract and thinks it can be made to work quite well.
The problem with supporting a paper-based workflow is that it will be masses of work for the organisation in question...