I don’t completely understand how things are evaluated here: when only some parts are present as open formats, the whole dataset is still indicated as open format? The reviewer explicitly says she confirms the content of bills are only available as html and pdf, but the whole dataset is marked as open and machine readable format. Does OKF now considers having a website and pdf files is sufficient requirement to be Open Data?
Also how come when some required elements like votes here are only partially published for some bills, it is still considered as OK?
Once again I feel like the method lacks in precision, saying things are fine when they are not is totally counter productive… Since multiple datasets are required per entry, the final result should reflect when things are only partial.
We don’t really understand this comment about pdf as data are available in JSON, XML, CSV or Postgre SQL 8.4 format. See as stated the relevant sites: http://data.assemblee-nationale.fr & http://data.senat.fr
I’m speaking of the content of bills which is part of the requirements : the actual text of a bill’s articles, these do not exist as official open data at all yet, there’s only citizen attempts at recreating it from the complex sources, see last parts of GitLaw: How The Law Factory turns the French parliamentary process into 300 version-controlled Open Data visualizations – Open Knowledge Foundation blog
Please excuse our belated reply. It is great that you highlighted the question of machine-readability and text data.
There has been a similar discussion in this thread where @herrmann left some remarks. Very short: I support your opinion @RouxRC, that text-based data needs to be published in a machine-readable format. @Slarrick wrote an enlightening piece as to why text-based data should be open too.
Do you know any best practices / prime examples of parliaments publishing legal code as machine-readable data? In my experience some governments publish info on draft legislation in XML - and nest a link to non-machine-readable bill content text in it.
Do we know of best practices how machine-readable bill content looks like? Ideally explained with some images? From our experiences, I know that it is very useful for government to see how a machine-readable bill dataset looks like - and it makes it much easier to point out flaws in non-machine-readable data.
Also, @RouxRC - you are right! Since bill content is one of the mandatory data elements of draft legislation, we cannot say that France’s draft is fully machine-readable. We followed this logic in our assessment and many apologies that this specific case slipped through. We’ll make sure to note this on the index page!
Hi @dannylammerhirt and thanks for the answer!
To my knowledge, the most advanced draft legislation publishing format is probably Akoma Ntoso : Akoma Ntoso - Wikipedia http://www.akomantoso.org/
I believe a couple african countries and Italy use it, but the best fitted to answer that question would be the people from the OpenParliament Declaration : https://www.openingparliament.org
Thanks for the very swift reply! Yes we have been in touch with Opening Parliament and the National Democratic Institute when defining this data category. I’ll dig again into their resources (as far as I remember the OGP Legal Openness Working Group also provides some best practices examples).
If I find something I will post it here and in a separate thread so it won’t get flooded in the (hundreds of) threads we already have. I will update the entry in France within this week.
In addition to the resources @RouxRC already mentioned, I’d like to point out some more:
As for examples of bills in machine readable format, legislation.gov.uk offers all of its content, both legislation and bills, in several formats. Actually I think it stands out as a gold standard for this. As an example, see this bill in XML format and the documentation for developers.
Maybe @ppkrauss knows more references and examples.