On 26 April 2010 18:43, stef <[email protected]> wrote: > hi, > > i just subscribed on advice by Tony Bowdens relayed by a common > acquaintenance. so here's my question: > > i ready you're also having trouble with diffing legal texts: > > http://www.guardian.co.uk/technology/blog/2010/mar/18/digital-economy-bill-parliament-diff
I think its true to say that a lot of My Society people wish it was easier to track changes in the development of Bills, not to say any legislations, which is the reason for the "Free My Bills" campaign. At the moment most of us use one or other relatively crude tool to work out how things differ. Francis Irving mashed something up when a group of us (not directly connected with My Society) were campaigning about the Legislative and Regulatory Reform Bill (as it then was) which worked on html, the Guardian seems to have done a diff on text extracted from a PDF (if I read it right). As it happens we do have an "easier" way to work this out for bills. A bill is printed at several stages through Parliament (representing the fact that it used to be expensive to do this - it originates at a time when amendments were scribbled onto a text and then written up later). Between those stages amendments are proposed in a form that explains in a formulaic (though English) fashion how those changes will affect the bill. In other words a "diff" written in Parliamentary language. A subset of those amendments are selected and it is those that get inserted into the next printing of the Bill. I think most of us are interested in capturing that level of detail because not only could we then automatically work out how Bills changed, it is at that level that votes are taken in Parliament and citizens can bring their pressure to bear on legislators. Keeping track of each individual change is also interesting as you can then trace back whose "fault" any particular provision might be 8-). Free My Bills will (hopefully) work that way but because there is a hope that we can persuade a sane Parliament to do the right thing and mark its bills up properly there's a reluctance to try to build a really cool tool to do the job in advance. > > we have this problem as well. we're analysing among other ACTA, the Gallo > report. the biggest effort is to clean the texts, so that the diff really > shows meaningful results. > > what's your solution? we tried the stock python difflib and a google > implementation of diff. I'm afraid I still just print the thing out and then mark it up with a pencil. -- Francis Davey _______________________________________________ Mailing list [email protected] Archive, settings, or unsubscribe: https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public
