Over in library land Oli announced the first release of his package for accessing the data platform. He’s made it available in two delicious flavours. There’s an R package which brings back data as tibbles. And a Python package that returns data as pandas DataFrames. Both connect to the public SPARQL endpoint we released back in September and provide a low level interface for SPARQL selects and a higher level interface for downloading specific datasets. If you do have a play with the packages, Oli is keen to hear feedback. If you want to tinker with the SPARQL endpoint, our Samu would be delighted to hear from you.
Anya, Ben, Michael and librarian Jayne went along to the Wilson room to hear all about treaties, brexit and the constitution. Arabella, Ewan and Eirik skipped through a brief history of the role of government and Parliament in foreign affairs, from the Ponsonby Rule to the Constitutional Reform and Governance Act 2010. A case was made for strengthening Parliament’s role, involving Parliament in the why and the what stages of treaty making in addition to the how stage of implementing legislation. Michael had some questions about how much the CRAG treaty procedure we’ve been sketching might be expected to change but was too shy to ask. This one hour talk shed more light on treaties and treaty procedure than we’ve managed to glean from two years of journalism. But that’s journalism for you.
Anya and Michael spent Monday morning wrangling House of Commons Jack’s edits into the procedure descriptions. They now await sign off from Jack and House of Lords Jane. With a little more wordsmithery the descriptions will hopefully poke out onto the website sometime soon.
Anya and Michael also made a second stab at cobbling together the logic and statements we’ll need to summarise the current state of work packages. They think they’re happy with the one for proposed negative statutory instruments but the rest still need work. They’re finding it’s all rather hard. Like code but with really, really long and slightly baffling variable names. So like code then.
Tuesday morning was supposed to be spent with Silver and the parliamentary archive elves continuing our explorations into ISAD(G) and Records in Contexts. But Silver copped a dose of the man-flu and none of the rest of them knew enough to continue. So instead, Robert and Michael took to the Tothill sofa and continued to carve away at comments. And Anya, librarian Liz and archive elf Nicky peeled off to the neighbouring sofa to chat about introducing Parliament’s concept vocabulary to the archives.
Anya, Robert and Michael took a stroll across the park to the plastic gothic Lego castle that houses the Parliamentary Computational Section. They met with Bex, Christine and Matt to chat about committee modelling and APIs and data. The boy Tetlow made a late but welcome appearance. It all went rather well. Plans were hatched. One feels many more meetings may happen.
Anya and Michael went to meet Bex, Christine, Alison, James and Usman to chat about subclasses and superclasses and URI structures. They have a problem to solve in that answering bodies, formal bodies, laying bodies and government departments have some overlap in function. They don’t want to give the same thing more than one URI so they need to group them somehow. But we lack a word to help here. Librarian Liz has been called upon. Thesauri will be thumbed. They also agreed the first iteration of these group pages will just be a list of pertinent links to subsidiary resources. “In the future we’ll know more and patterns will emerge,” said Christine. And she’s not wrong here.
A little while back we reported that Anya had been chatting to the kindly souls of the Procedural Publishing Unit (PPU) about adding anchor links to Future Business papers. All ably assisted by our Jake’s proof of concept browser extension. Those links are live which means we, or indeed anybody else, can link directly to an item of future business rather than to the whole of future business. PPU Mark has since done the same for Votes and Proceedings and next Monday’s Order Paper is the first to ever contain links to business items. This really is all quite splendid. Better links making a better web and all that. Thanks Mark.
Michael spent a train journey adding more OWL imports to our interface classes. About which he is quite ridiculously excited. So far we’re covering the agency, business item, laying, legislation, making available and procedure models with a dash of SKOS thrown in for good measure. There’s a hand drawn picture here and a machine generated version here. Beautiful. Absolutely beautiful. As we’re sure you’ll agree.
In the course of making interface classes, Michael also made a first stab at the making available model. It’s a way to wrap up the assorted routes by which materials are made available to a House or to a House via its library. In the future it will provide a means to aggregate all the things that have been made available to a House by a body at URLs like groups/yiuTaSI4/made-available. As ever, it needs some Anya and Robert time to hammer out better comments.
Robert and Michael continued to hone model comments to a fine point of precision. It’s like Hemingway meets Tulkinghorn on the Tothill Street sofa. Though Charters and Caldicott might also be a fair analogy. Some say.
Last week our librarians and clerks were stretching their heads around questions of parliamentary time. When a sitting overshoots the midnight hour, Parliament considers the current day to be the previous one. This raised the question of whether an affirmative statutory instrument being approved in the early hours would become law on the calendar day or the parliamentary day. John got back to say he believed “the law follows Parliamentary date when it comes to questions about what day SIs are laid / approved etc, rather than actual date.”. He then passed us on to Rich for confirmation. Which didn’t quite come. Rich told us:
Can’t find a more authoritative source than JLS, and granted that SIs can be laid in office hours or until rise of House, whichever is later (SI Practice ¶4.8.10ff), but … I’d have thought statutory timing requires statutory or common law (not parly) interpretation. […] thus laying days would be delimited by midnight, not by adjournment. (Presumably the quirkier rules apply for internal parliamentary purposes, such as intervals between bill stages.) SI Act 1946 and Laying of Docs Act 1948 are silent on “day”. Certainly, sitting beyond the start of the next day’s planned sitting forces abandonment of the latter sitting (Erskine May, 21st ed, p239). Consolatory trivia: “The longest sitting on record occurred on 31 January 1881 on the Coercion Bill, when [HoC] sat for 41½ hours.” (ibid)
That the law might be silent on the definition of “day” is a thing of rare beauty. Some people say arcane, but so is poetry. On the subject of Parliament and time, that the “The Calendar (New Style) Act 1750 makes specific provision for the year 2800 not to be a leap year” is also a marvellous new fact.
Friday morning saw Anya, Robert and Michael reunited. This time with Mark Fawcett and Matt Brown from the Procedural Publishing Unit . They took a first stab at modelling House of Commons Standing Orders. It’s a bit of a side project to turn some fairly flat HTML into something more weblike and linky. Several iterations of whiteboard and marker later a picture emerged that everyone seemed most pleased with. Top work team.
The mighty Samu fulfilled a promise and added IPTC compliant metadata to our lovely photos of Members. Which means Google images now carries a credit for the photographer Chris McAndrew and an even more exciting Creative Commons Attribution licence. As Jamie puts it, some “proper asset management with standards compliant data leaking out all over the web - more lovely work from @langsamu, transparent and open as ever.”
Alongside a slew of other open source tools we both use and contribute to, we often call upon WebVOWL for both idiot checking and communicating our work. Throw an RDF data model in its general direction and it does all the hard work to parse and visualise it. Which is all rather handy. But even when nicely laid out, the full physical ontology can be a little daunting. So Samu wrote a new endpoint on our Query API that gathers a subset of the full ontology. In this case it’s grabbing the bits of the model pertinent to Michael’s recent post on attempting to teach parliamentary procedure to machines. You can see it visualised in all its glory here. Nice work as ever, Samu.
…once again has joint winners. Step forward Chris and Mike.
In tune with the times, Dan was approached by some people on Twitter to ask if we had a list of Members of the House of Commons who also hold paid government positions. In some ideal world, Parliament would publish a register of Members, government would publish a register of paid positions and incumbents, identifiers would align and distributed queries across liminal zones would make all things simple. But we do not live in an ideal world. So we have to rely on the data we store about government positions. Any road, Chris got to work with his trusty SPARQL spanner and came up with this beauty. Look at the sheer length of that monster. We’re gonna run out of URI pretty soon. The number returned - 96 - is somewhat short of the figure people have been quoting so it’s possible that our Members’ Names Information System definition of a government job is different to the definition used elsewhere. Nevertheless, it proves we can use our public SPARQL endpoint to write and share these sort of queries rapidly, which we hope will encourage people to have a tinker.
Our joint prize winner Mike has been helping out Oliver Heath, a professor of politics at Royal Holloway. Oliver’s working on an ESRC funded project to map census data onto constituency boundaries from 1851 to 2018. He plans to investigate how the social correlates of turnout, electoral competition and representation have changed over time. The European Commission are partners in the project and we’ve also been asked to join. Mike sent him links to our APIs for constituency data - amongst other things - and a link to the election model. We’re all hoping for exciting things here. Top work Mike.
It is exactly one month since Alex triggered his personal Article 50 so, at five on Friday, he collected his parliamentary mug and teddy bear collection, handed in his pass and disappeared over the horizon. At least as far as the Red Lion. For pastures new, fame and fortune. Or comparative fortune at least. Best of luck Alex. It’s been a pleasure.
Anya, Ben and Michael strolled as far as the Strangers’ Bar. But the cold made even that short trek a struggle.
Chris took off on a fact finding mission to Hungary. We’re told the goulash they serve in the parliamentary pudding section has practically nothing in common with actual gulyás. We’re also told that Halászlé is amazing. And the mountain cable cars are amazing. And lángos are amazing. We also learn that Kerepesi Cemetery is not the same thing as the cemetery in Kerepes, and confusing the two can waste a lot of time and energy.