Transcripts of conversations with mySociety people

Here the right answer is “it depends”. There are many factors, beside the skills of people who are writing the parsing code, that come to play. It depends on how well or bad the document that you want to parse from is structured, and how uniform those structures are. It depends how semantically rich you want the Akoma Ntoso output to be.

Parsing legal documents is not just a coding exercise, it requires a good understanding of the structure and semanticsof the document. Usually transcripts do not have a complex structure, but in the case of bills and acts, or judgements, to create a good parser you also need the input of people with legal backgrounds.

I have to say that Akoma Ntoso may be a bit harder to digest, at least at first, not so much because Akoma Ntoso is complex, but because it is different.

Yes, when you open the Akoma Ntoso “box” you don't find, say, a debate schema, prefabricated for you.

Akoma Ntoso is more like a Lego box full of building bricks. Yes, you do not have the prefabricated house waiting for you, but if you only have the patience to learn how to put together the bricks you have as many "houses" as you wish.

Like Lego, you can build very simple houses with just four walls and a roof, or you can build elaborate castles. Akoma Ntoso can be as simple or as complex as you wish and it's just about as intimidating as Lego box of bricks. Just play with it!

Keyboard shortcuts

j previous speech k next speech