For Anya and Michael the week kicked off with a Monday morning meeting with library statisticians Cassie and Oli, and the ODI’s very own Leigh Dodds. And continued into the afternoon with Leigh and Samu. The House of Commons Library takes in a variety of statistical data from a variety of sources and presents it as a set of constituency dashboards. Primarily for the benefit of Members and their staff but online for anyone who may be interested. The data is currently stored as a set of spreadsheets and we’d like to bring it into the data platform, link it up with everything else and enable people to ask questions like, ‘do Members representing constituencies with some demographic characteristic tend to ask more or less questions on subjects associated with that demographic characteristic’. For instance. A few months back Anya, Silver and Michael did some work on a stats series model which they thought covered most of what was needed. They also took a look at RDF data cube but their little brains refused to map the examples to the schema. So they asked Leigh for a borrow of his brain. Leigh is currently typing up a report which they hope will provide some advice.
On Tuesday, Alice from the Institute for Government crossed St James’s Park to visit Anya and Michael in library land. Alice co-authored the Parliamentary Monitor report which came out earlier this year. Anya and Michael explained some of what they’re doing with domain modelling and some of the reasons for doing it. They had a good conversation about the problems people have working with parliamentary data, particularly the amount of time and effort it takes to burrow through large documents to get to the data. Alice is a big fan of Sessional Returns, but an annual publication can’t meet the need for ready access to recent data. Alice is planning to write next year’s Parliamentary Monitor; hopefully we can make her life slightly easier before then.
On Wednesday, Anya, Robert, Samu, Ben and Michael called by a committee room to see a demo of the new Bills authoring software which is some kind of collaborative project between UK Parliament, Scottish Parliament, The National Archives and some other people I’ve forgotten. It’s based on Akoma Ntoso and there’s not really an awful lot else to say. It was kind of impenetrable.
On Thursday, Michael’s old boss Tristan came over to library land for a chat with super clerk Gordon Clarke. Gordon wrote up some thoughts on what he thinks needs to change about committee reports in the not too distant future. Some of it touched on fragmenting documents, some of it was more about constructing documents from fragments, some of it was in the general area of reproducible research and the kind of things Tony has been looking at. And some of it was about variable length, variable depth, and variable audiences. Tristan and the BBC R&D team have been doing work around variable length content and all things stretch texty so Michael set up a chat. It was good. And lovely to see Tristan again.
Up on the 5th floor of Tothill Street, librarian Jayne is continuing in her role as the Wizard of Oz of Statutory Instrument tracking. She spends her days combing through long documents looking to see what’s changed, adding business items and actualising steps. One problem Jayne faces is that Parliament has a tendency to take massive documents designed for printing and stick them on the web. Which means Jayne can link to roughly, but not exactly, the right place. Not ideal, given the complexity of parliamentary procedure. Anya has managed to persuade the people who publish the Future Business Paper to add some IDs to their headings. Which might seem a small thing to the rest of the world, but is quite a breakthrough for us. It means Jayne can now link to the bit of the document that covers the business item, rather than to the top of a huge document and hope users know how to Control-F. Unfortunately, the IDs they’ve added are not yet exposed either inline or via a table of contents. So our Jake has written a JavaScript bookmarklet to expose IDs and generate a table of contents. We hope this makes Jayne’s life easier. We also hope this may find its way into Future Business proper. One day.
In another exercise in ‘reaching out’, Michael talked for two hours to assorted librarians and computational staff about the various models involved in the tracking of SIs. Michael was feeling a little below par, and fears neither his brain nor his heart were really in it. But he thinks he got away with things. Mostly. As a side benefit he managed to shunt a few extra models into the interface descriptions so the big picture of how everything clips together is coming along rather nicely.
Anya, Robert and Michael made a rare excursion from their library sofa to go visit Samu in the Parliamentary Computational Section. The subject for discussion was the addition of logic gates to the procedure model. Which has been raging on for a little while. Both Samu and Robert seem to agree that having both route types and logic gates is overkill. Michael’s brain won’t quite expand to whatever it is that that entails. A still small voice says some things cause things, some things allow things, some things preclude things and sometimes all of that is conditional on something else. But the general consensus was that the addition of logic gates would tighten up the model and make it more descriptive of procedure. So that’s good. More thinking and scribbling needed here.
Michael had another brief catch up with House of Commons James and House of Lords Edward to dive back into committees and their types and the type of things they scrutinise. They chatted about the shape of the API we’ll need for expressing type information. And how we might need a many to many relationship for describing scrutiny areas. Another meeting is planned to chat about the URI patterns we’ll need for this.
Chris and Samu put their heads together to write some code, taking a ‘difference report’ and allowing it to be easily turned into a SPARQL query that applies any changes found. Which means we can keep the data in our triple store perfectly in sync with its Git repo, but without the current process of deleting and replacing it in its entirety. The old approach lead to all our inferred triples needing to be re-inferenced every time a change happened. Which had some pretty terrible performance implications. They’ve submitted this as a patch to dotNetRDF, a library we both rely on and contribute to. Top work Chris and Samu.
Chris also spent some time building a class explorer for the ontology in our triplestore. Unlike relational databases, triplestores allow an item to belong to many classes simultaneously which allows us to categorise things by their behaviour and relationships rather than by more rigid table membership. Chris’s new tool, in showing how many members a class has in common with other classes, allows users to understand both the quantity of available data by class and how closely related classes might be. Chris would like to add that the CSS is something of a ‘mare.
Alex has continued on his mission to expand the coverage of search hints by editing the existing rules. He says it’s going alright but Parliament has lots of truly terrible URLs. And he’s not wrong.
Not so much a stroll as a train ride followed by a short hop on a bus, as Anya, Robert and Michael travelled to the Wayzgoose Book Fair in Oxford Town. Not on a working day, I hasten to add. Working days are for noses. And grindstones. In the market hall, Anya and Robert paused to admire some poppy based baking. Then off to the book fair. Michael bought himself a poster from Red Plate Press which now adorns the walls of his office. Very swanky. A rejuvenating stroll around Oxford Botantic Garden followed. It is, quite frankly, neither as big nor as interesting as the one in Cambridge. For fans of botanical gardens out there.
Anya, Robert, Ben and Michael took a second outing on Tuesday night. This time across London to the Miller where Al Robertson told them all about the occult roots of modern computing. Michael tried to fit in with his new fortean friends by stealing Ben’s hat. Which is a change in circumstance, cos that’s usually Robert’s job.
“[J]ust as Nietzsche analysed the imposition of morality on the strong by the weak as an expression of ressentiment or a form of revenge, so metrics – the moral code of a sourly reductive managerial culture – are the means to make sure that professionals’ working conditions should more and more correspond to the alienated, insecure, hollowed-out working conditions of so many other members of society. [..] it is hard to see what might lead to the overthrow of the ‘tyranny of metrics’. The dominant managerial culture is not about to surrender its most cherished weapons, and somehow I don’t see the great uncounted storming the Bastille of performance data to demand the release of human activities imprisoned in graphs and tables. Critique may do something to weaken the superstitious faith in the omnipotence of numbers, and we may even come to accept that not everything that matters to us can be measured. But the desperate yearning for unobtainable forms of precision and objectivity may be too strong even for those modest triumphs.” Kept Alive for Thirty Days [LRB - subscriber only].
A deepfake news anchor that’s basically the same as any other news anchor
Michael spent a chunk of the week listening to Yeah I’m on Acid by the popular beat combo Shit and Shine. Recommended. Marvellous stuff.