Once more, our poor weeknotes have slowed into a monthly rhythm. Or ‘cadence’, as Young Robert would doubtless say. And once more we apologise. This time, we’re blaming indolence and Michael’s seemingly endless vacations. If we don’t pull our finger out soon, we could easily find ourselves publishing the worst of all things - ‘quarterly updates’. And I don’t think any of us want that.
Anya and Michael have given some thought to retitling their output, ‘month notes’. But a change to the title would also mean a change to the URI. And if there’s one thing cool URIs don’t do, it’s change. What don’t cool URIs do, guys? They do not change. And with that quick burst of polemic and a sarcastic wiggle of our eyebrows disposed of, we move on. It’s time for the ‘content’.
First up, we’re delighted to welcome Librarian Susannah to our crack team. Susannah is returning to the fold having been part of our team some years ago, before departing to lead the team responsible for acquisitions, disposals, journal purchasing and whatnot. We hope she enjoys her stint back on the more procedural, pedantic web side of Library life.
Susannah’s task is to pick up work we attempted to kick off last year - it could have been the year before, one year being much like another, except, perhaps, warmer - with the working title of, ‘a single subject view of the Library’. Or is it, ‘an information architecture for Research and Information’? Given this is in part about taxonomies, you’d think we could settle on a preferred term. To explain: despite our crack team of librarians pouring all their love and care into maintaining Parliament’s Thesauri, much of the software commissioned by the Library makes no use of their efforts. This is not because the taxonomy is lacking, but because the IT plumbing required to make use of data managed in one system in other systems doesn’t exist and has seemingly never been thought of. Never commission black boxes being a lesson one might learn. Though let’s not get into ‘systems thinking’ here. Those people are weird.
This means that the Library’s book catalogue uses a frozen copy of the taxonomy-as-text-strings, the website uses a list of Wordpress ‘tags’, the enquiry system lacks any subject descriptions whatsoever, the research briefing application has its own set of hard coded ‘topical’ terms and subject access to the subject specialist directory is managed in a creaking SQL database using ‘some words’.
Susannah has begun by returning to the work undertaken last year by Silver and his Data Language colleagues, to rebuild the subject specialist directory on top of our taxonomy, and using that work to propose and document patterns for future integrations. But subject access is merely one angle. At some point, we’d like to expand out to a full redesign of the information architecture of the Library. Or a knowledge graph if you will. Très moderne.
Last time out, we reported that our Jianhan had done all the preparation work required to integrate our beloved egg timer with our procedure editor application. New database tables were added, new forms built and existing forms fettled. Since then, he’s also updated our ontology to cope with all things clock related, fine-tuned our orchestration code to slurp clocks from database to triple store and written a new script to call the egg timer and reset our procedure editor clocks. All of this in our staging environment.
Librarian Jayne having tested and declared herself more than happy, Jianhan did all the same work again. This time in live. And by gosh and by jingo, it all works. Well, aside from a side kerfuffle with British Summer Time, which is only to be expected. We’re old hands at this stuff; the day we don’t trip up over British Summer Time is the day we hang up our boots.
All of which means that the next time sitting days change unexpectedly, our crack team of librarians no longer have a hundred and odd work packages to update by hand. Having ticked off her many and varied egg timer integration tasks, so pleased was Jayne, she put forward the offer of drinks to both Jianhan and Michael. We’re not sure she knows what she’s getting into here. Our Jianhan is not a massive drinker but Michael drinks like the proverbial fish. We hope she comes equipped with a credit card the next time a pub trip is mentioned.
Given this is - hopefully - the last time we’ll mention the egg timer for a wee while, a little background colour may be necessary. Parliament having passed a bill and the bill having gained Royal Assent, it becomes an Act of Parliament. Acts of Parliament may - and often do - delegate powers to ministers. Such powers are often accompanied by duties and those duties are often toward Parliament. Sometimes those duties are as simple as the laying of a paper. Sometimes, they’re slightly more onerous, setting out that a paper should not only be laid, but should be subject to some form of parliamentary procedure. Where a paper is laid subject to procedure, legislation will set out how long the paper should sit before Parliament. In calculating this period, there are three factors ‘in play’:
the date from which the calculation should start counting - being either the date of laying, the date of making or - in the case of published drafts under the European Union (Withdrawal) Act 2018 - the date of publication,
the number of days to count,
the type of days to count.
The last factor is where things get complicated. Sometimes - for published drafts, proposed negative statutory instruments and treaties laid under CRaG we’re told to count only joint sitting days. And sometimes, we’re told to count joint sitting days and periods when both Houses are adjourned for less than five days. At other times, we’re told to count joint sitting days and periods when either House is adjourned for less than five days. And for Commons only instruments, we’re told to count Commons-only sitting days except when the Commons adjourns for more than four days.
Quite why so many flavours of days to be counted exist, we have no idea. We’ve sometimes suspected a copy and paste error on the part of legislative drafters. But that can’t be right. If you know why the clock sometimes stops when both Houses are absent and sometimes stops when either House is absent, please do get in touch. We love a good story.
Until COVID hit us all, the job of calculating such periods was in the trusted hands of JO Jane, a slide rule and an abacus. Then lockdown happened and Michael figured he could automate the job in a couple of weeks or so. Some months later, our beloved egg timer went live. Which at least appeared to please Jane.
Sadly, the pain did not end there. Given the calculation styles are based on sitting day patterns, given sitting days are wont to change, and given we often have upwards of 100 instruments before Parliament at any one time, every late breaking sitting day news item caused librarians to groan as 100 odd clocks required updating. Jianhan’s fine work now fixes this. Our crack team of librarians update a set of Google calendars with assorted flavours of parliamentary days, a script takes this data and populates the egg timer and a new script running in procedure editor checks the egg timer every 10 minutes to reset clock end dates as appropriate. Lovely stuff.
Having completed the egg timer integration, thoughts have turned to a couple of requests from JO Jane. Namely, how to distinguish between current and defunct laying bodies on our statutory instrument website and between current and defunct lead organisations on our treaty tracking website, how to distinguish between current and defunct procedures, and how to order procedures in a fashion that does not involve bastardising their names. The bastardisation of procedure names being our current - somewhat embarrassing - approach.
Having planned out the work, Jayne and Michael came to the conclusion that the data to distinguish between current and defunct bodies was already in the triple store and that making that distinction visible to users lay entirely in Jayne’s more than capable hands. Tickets were created to add appropriate queries to our ever-growing SPARQL library. At which point, Jayne found she’d already made such a page. Always nice to find you’ve already done the work you’d planned to do. Current Jayne would like to thank past Jayne for her diligence. More tickets have been created for our colleagues in Software Engineering to take Jayne’s queries and do what they need to do at their end.
Distinguishing between current and defunct procedures - and the ordering thereof - remains new work. And work that can only be performed by computational single-point-of-success Jianhan. The ‘Jianhan doing’ column on our machines Trello contains all the tasks necessary at our end. Should you be interested. Which you probably won’t be. If Jianhan ever gets any time - and at the moment that seems unlikely - he hopes to slot in the work over recess. Poor Jianhan.
General election planning took a significant stride forward this week, when Sym reported back on progress adding MNIS Member IDs to Democracy Club candidates. Not only is the new field added, it is also populated. This made much easier because Wikidata stores both Democracy Club candidate identifiers and MNIS Member identifiers. So Sym was able to take a Democracy Club candidate ID and ask Wikidata for its MNIS Member ID equivalent. Nifty. Like a modern day Rosetta Stone, if you will. We still have a job to spot test the results, which we’ll hopefully pick up next week. Thanks Sym.
In other election news, Librarian Emily and computational bazballer Michael feel they’re finally getting somewhere with their understanding of when and why a Member of the House of Commons might cease to be a Member of the House of Commons. Michael spent part of his otherwise monumentally unproductive week redrafting their end reasons list and end reasons diagram to reflect a more legislative stance. A thing he’s secretly quite pleased with. Yet another meeting with researcher Neil is pencilled in for next week - sorry Neil! - at which point they hope to finally put the problem to bed and data tidying can commence. Top work Emily.
The people side of things largely taken care of, attention has turned back toward places and parties. New constituency names finally having been finalised - ish - researcher Carl sent through a spreadsheet, which Librarian Anna has incorporated into ours.
On the party front, names, abbreviations and initials are all now tidied and our party name policy fully documented. Fine work team:Phil.
Librarian Jayne and computational midshipman Michael have finally finished decanting all of Matt and Mike’s comments into our motion cheat sheet documentation. So, if you’ve ever wondered what conditions need to exist before a given type of motion can be tabled, debated or have its question put, we think we have you covered. At least for statutory instruments. Thanks, as ever, for your time and patience Matt and Mike.
In a change to usual procedure - and indeed precedence - Librarian Jayne has been removing steps from our procedure maps. Well, one step, in one map. That step being the opening of consultations by the government on free trade agreements. Since we already have a step describing ministerial statements announcing the consultation being opened, we came to the conclusion that a step describing the consultation itself was not only redundant but not particularly parliamentary. So that step is now gone.
We’re delighted to announce we’ve finally been joined by an actual developer. Incredible news. Will wonders never cease? Developer Jon has joined our crack team of librarians, computational ‘experts’ and designer Graeme to take on the job of turning our shonky search prototype code into something we can actually work with. And indeed test. Before Jon joined us - or got himself ‘onboarded’, as we like to say in these parts - computational leg spin bowlers Young Robert and Michael spent a little time refactoring their search prototype code, swapping out their generic object rendering template for something based more on the class of the object being rendered. Jon’s next job is making our fat views and controllers thinner and our nonexistent models considerably fatter. At which point, we’ll have a reasonable starting point for an actual test suite. Welcome aboard, Jon.
It goes without saying that Jon’s new search code is of little use if there’s nothing to search. To that end, our Jianhan has been trying to wrap his considerable brain around the problem of porting our antiquated Solr index to something more modern. Or at least to some version of Solr that’s still actually supported. Poor, poor Jianhan. At some point, we plan to plug Jon’s front end code into Jianhan’s refreshed Solr, turn off our search mock prototype and save boss boss ‘brarian Bryn’s credit card a dollar or two. Though that prospect seems to be disappearing down the road and possibly over the horizon.
Design meetings with Graeme also continue apace alongside the usual project ‘check-ins’, conducted in new-fangled software called Miro or Jiro or Giro or some such. In the prime of their lives they may be, and Young Robert and Michael begin to suspect this was not designed for eyes of their age. Especially when seen through a laptop, through a Teams window, through a not quite fullscreen browser window. Purchase orders for monocles and magnifying glasses have been raised.
Work progresses to design and document models for what we continue to hope will be our new new search. That’s the one based on new data ingests, new data models, new data transforms and entirely new search stack. Or so we hope.
This time out, we’re happy to report that two new models have moved over to our done pile. The oral contribution model fits the Ronseal definition, describing credited and uncredited oral contributions to debates. For some definition of debate. The paper type model builds atop the paper model - and its inbuilt taxonomy - to describe some of the relationships between assorted types of paper. Lovely stuff.
Week 28 saw a minor triumph on the open data front. And not one of ours. Researcher Elise is knee-deep in a machine learning project aiming to track - and group - bill clauses and amendments over time. In the usual parliamentary fashion, the numbering and citation of clauses depends upon positioning, which changes over time. See our whining on standing orders passim.
This obviously makes analysis and comparison more difficult than one might like. Elise emailed Anya, who pondered aloud whether bills authored in our new Lawmaker system might have persistent clause identifiers and suggested Elise might want to contact the Legislation Office to see what might be made available. All of which led to the first - to the best of our knowledge - publication of Lawmaker XML for a bill. In this case, for one expression of the Digital Markets, Competition and Consumers Bill. More excitingly still, the publication came with a commitment to do the same for all future expressions of Lawmaker authored bills. Michael could barely contain his excitement as bill XML documents started to drip into our bill paper application. Top work all round.
Still with Elise, she got in touch recently - well, fairly recently, given current ‘cadence’ - with an enquiry about the number of Members in the House of Lords at various points in time. Unfortunately, this coincided with vacation season for Librarian Anna and Librarian Emily. More unfortunately still - for us, at least - it also coincided with Librarian Phil’s paternity leave. At which point, we break off to offer congratulations to Karunja, Phil and family.
Luckily House of Lords Library Henry was present and correct and more than able to rescue Anya and Michael from the hole they’d managed to dig for themselves. Thanks Henry.
Hot on the heels of Librarian Jayne’s efforts, boss ‘brarian Anya has published her very own update to a Facts and Figures briefing, this one on addresses to both Houses of Parliament, gaining her very own Library author page in the process. It’s the first spreadsheet in the Facts and Figures series to feature our new ‘about’ tab. Or ‘metadata textbox’ as Michael likes to call it. An innovation invented by Anya and Michael over a pint of Guinness on the Strangers’ Bar terrace. They just don’t know when to switch off, those two.
The second in our hopefully ongoing Wikidata community meet-ups took place back in week 27. This time with the Commons Library hosting. Attendance was somewhat down on last time, with Dan unable to attend and Andy finding himself on the receiving end of a pretty terrifying looking promotion. Still, alongside everyone’s favourite team of crack librarians and computational oddjobsmen, Jeremy and colleagues were there to rep the BBC, Paul and colleague joined from Levelling Up land, Silver and Louie were reppin the south coast massive and Andrew was on hand to supply actual, functional advice. Because of course he was. We think everyone had a fine time, if not at the actual meeting, at least in the pub afterwards. Our next outing is not until September, when we hope to visit Jeremy in Broadcasting House. Getting Michael to leave again may prove tricky.
It’s not often one sees one’s work described as genius. To the best of Michael’s knowledge, it’s only happened maybe a dozen times in his educational and working life. Imagine then our surprise, when we clicked on a link in Twitter and found these weeknotes - the very ones you’re reading - described as just that.
It came at a critical juncture for your regular correspondents, who’d been pondering whether to stick with their current style or pivot toward the more professional. Seeing how the big boys handle ‘comms’ matters having undermined their confidence, they’d been toying with the idea of delivering updates through the medium of non-tabular data displayed in tabular form, in a single slide ‘deck’, open in edit mode, and distributed via Teams. A poll was organised to gauge the temperature of our dear reader, and we’re delighted to say the consensus seemed quite settled on sticking with our usual blather. Which is nice.
Anya and Michael would like to thank Giles and will do as soon as they put down their copies of the Cluetrain Manifesto and wipe these blasted tears from their eyes.
