Week 17 saw our much coveted Librarian of the Week trophy once more adorning the mantlepiece of a non-librarian. It will surprise almost no one to learn that our Jianhan was the non-librarian in question. If you’ve been following along from home, you’ll be well aware that our major plot line of late has followed Jianhan’s attempts to upgrade our aged triplestore. Last time out, we reported that his efforts had been signed off by our crack team of librarians and pushed live.
The upgrade happened - as our major upgrades usually do - during recess. Or, more correctly, recesses, the two Houses ploughing their own furrows on such matters. The following week, our Members boarded the Hogwart’s Express bound for Westminster, the inevitable deluge of digital data following in their wake. The first day of term has always been particularly busy for our crack team of librarians, floodgates opening as upwards of a thousand written questions submitted during recess get their ‘tabled’ button pressed. “All hands on deck,” Librarian Anya has been heard to shout, anywhere up to 14 librarians putting in a nine hour shift. The more librarians joined the fray, the slower the software got. But not this time.
Dear reader, Jianhan’s efforts worked. Not only did they work, they worked damned quickly. “It goes like sht off a shv*l,” a librarian exclaimed. Not words one usually associates with a librarian. We worry they might be spending too much time with Michael. Notwithstanding this slight coarsening of language, our reporter was not wrong. All tabled questions were subject indexed before the afternoon tea trolley arrived. Remarkable.
The upgraded triplestore comes with a copy of Workbench layered over the top of its SPARQL endpoint. Which means, for the first time, our crack team of librarians can get their eyeballs on the data they’re tasked with managing. It is a managerial cliché to meet efficiency gains with, “this is not taking away your jobs, this is freeing you up to be more creative”, only to walk into an empty office a couple of weeks later. But clichés are clichés because they’re occasionally true. Librarian time spent waiting for the blasted computers to save things, can now be spent tending our data infrastructure. Now all they need are the tools to fix the data without having to telephone the Parliamentary Computational Section. A development we hope to plan for soon.
Well done Jianhan. We’ve rarely seen our librarians looking quite so happy.
With the triplestore upgrade out of the way and Librarian Jayne back on duty, Jianhan has also found time to attend to a bug or two in the pipes between our procedure editor database and yet another of our triplestores.
A step we dearly wished to delete stubbornly refused to comply with our demands, whilst a handful of business items had failed to show up in the triplestore. Eventual consistency, they said. Such is computational life. Anyway, they are both now fixed. More thanks Jianhan.
Also fixed is a small bug whereby business items describing laying in the House of Lords were not being populated for bicameral statutory instruments. It turned out the ordering of identifiers was important here. As the ordering of identifiers so often is.
If our dear reader tuned in last time out, they’ll know that Librarian Ayesha and her computational helpmate Michael got themselves in a proper tizzy over the legislative conjugation of ‘laying’ and ‘statement’. Not words a parliamentary pedant expects to find in the same sentence. Or at least not these parliamentary pedants. We’d failed to properly read the legislation, seeing only the “statement”, and skipping quite past the “laying” of it. Confusion was compounded by the making of a statement following the laying of a statement for the Agreement, done at London on 14 June 2023, between the Government of the United Kingdom of Great Britain and Northern Ireland and the Swiss Confederation on Recognition of Professional Qualifications, whilst the One Hundred Year Partnership Agreement between the United Kingdom of Great Britain and Northern Ireland and Ukraine saw no such effort made. Our treaty procedure map has now been updated to reflect our new understanding, steps being added for the laying of a statement in both the Commons and the Lords. These new steps are accompanied by an optional step describing any subsequent making of a statement. Again in both the Commons and the Lords. So that’s cleared that up. Splendid.
Computational pipe cleaning and cartographic exploits all help pass the time, but would be for naught - or nowt, as Michael might say - if eyeballs never met pixels. To that end we’ve also been hard at work on our Procedure Browsable Space™, which this week sees the addition of a new page listing work packages containing business items actualising one or more steps in a step collection. Quite the mouthful. Or gobful, as Michael might say. Here, for instance, is a list of work package where at least one committee has raised a concern. For those who’d like an alert whenever there’s a new entry on this illustrious list, an RSS feed is available. Because of course it is. Very handy as we’re sure our reader would agree.
Also joining the done pile are pages listing work packageable things and work packages for work packageable things, as well as pages for preceding work packageable things and following work packageable things. At which point, our dear reader is most probably asking, “what on earth is ‘work packageable thing’? Have you gone quite mad?” Well, a work packageable thing is the focus of a parliamentary work package. It describes a set of things including bills, secondary legislation and treaties. Whilst bills and secondary legislation might both be described as legislation, treaties aren’t. And whilst secondary legislation and treaties might both be described as instruments - or indeed papers - bills aren’t. Let’s not even get started on FRBR. So, for now, we’re stuck with the rather ungainly ‘work packageable thing’. If you have a better suggestion, we’d be eternally grateful. Answers on a postcard, please.
Computational Section colleagues have made a couple of welcome changes to our treaty tracking website, bringing URLs for work packageable things and work packages - yes - in line with our statutory instrument website. Consistency not being nothing. The same release also saw a dash of consistency applied to the display of business items actualising more than one step, as can be seen on the timeline for the One Hundred Year Partnership Agreement between the United Kingdom of Great Britain and Northern Ireland and Ukraine. 25 March saw the International Agreements Committee publish a report not only drawing the treaty to the special attention of the House, but also calling for a debate on the matter. Splendid stuff.
In the continuing absence of a developer, work on new, old search has ground to a halt. Poor Lydia continues to jump through recruitment process hoops, only to find more hoops spawning in their place. Like Sonic the Hedgehog, if Sonic the Hedgehog were somewhere between exasperated and baffled. We do, however, have one thing to report. Both versions of our Poller code - the bit between the triplestore and Solr - now make use of the external MNIS API, rather than the internal one that had been in use. We suspect that this actually happened months ago, but someone forgot to move the ticket. Included here then for completeness and as a reminder to self to keep moving the tickets. Else we’ll never know what’s done. And what isn’t. Can someone create a ticket for that?
Psephological efforts remained focussed on importing our Parliamentary facts and figures list of maiden speeches to the election results website. A simple task in theory, made more difficult by many and varied mismatches between constituency identifiers and labels across assorted systems and time periods. Which means we’ve been mainly decanting data between systems by means of a slightly leaky bucket.
Librarian Phil has been triple checking that the 1997 constituency names stored in MNIS match the legislation that created them. Over in election results land Michael has also been engaged in somewhat similar work, this time to align 2005-10 constituency names with the legislation. He’s also stripped out a rather annoying “smart quote” that had somehow crept into Regent’s Park and Kensington North. Never let Microsoft get involved in your toolchain. That, then, is names sorted out.
When the 1997 constituency data came through from Statistician Carl, it was accompanied by MNIS constituency identifiers. Damned handy as it turned out, but not something that had been present in the spreadsheets covering the 2005-10 and 2024 constituencies. We’d previously been keying constituencies off ONS issued geographic codes and such codes were not minted for the 1997 constituencies in Northern Ireland and Scotland. So getting MNIS identifiers for all constituencies became something of a priority. Librarian Phil provided a complete list, which Michael has now imported into psephologyland. Lovely stuff.
Meanwhile, work was underway on the maiden speech PFF. Librarian Anna added ONS issued geographic codes across all pertinent constituencies from 1997 onwards, whilst Librarian Emily has been busy adding MNIS constituency identifiers to the same spreadsheet. Which means MNIS, election results and the PFF are finally in sync for both constituency names and constituency identifiers. Michael has, in turn, taken that spreadsheet and rewritten the script to load maiden speeches into the election results database. Keen eyes will notice that that script only imports a maiden speech if it can match on both the Member and the constituency the Member represented at the time the maiden speech was made.
A few template changes later and links to maiden speeches now pop up on Member pages, constituency pages and Parliament period pages. As our constituency data only stretches back to 1997, maiden speeches made by Members prior to that are not yet linked to. For which we can only apologise. Rest assured, matters will improve as our historical constituency data improves. An awful lot of below the surface paddling for a small splash of hypertext. But that’s librarian life.
A couple of other additions to the election results website are links from constituency pages to two new Commons Library data dashboards: one describing voting patterns across different demographics, the other sewage discharge data in England.
Still with our lovely little collection of Parliamentary Facts and Figures, this week saw another publication emerge from the IDMS paint shop, this one listing living former Members of the House of Commons. If you’re in the market for non-dead former Members with dates for first entry into the House, dates of departure, and party affiliations at the point of departure, this is your one stop shop.
Our crack team of Commons Member data managers have noted our MNIS database records whether a minister is paid or not as a boolean on a Member’s incumbency in their ministerial position. Which means, should a ministerial position flip from being unpaid to paid - or indeed vice versa - we need to close their existing incumbency and create a new one. All of which makes incumbency data harder to query, which is less than ideal. Without a change to the MNIS database, there’s nothing we can really do to mitigate this. Computational Section colleagues have been alerted.
In other late breaking, administrivia news, deep within the depths of MNIS there is a field storing Members’ preferred names. Whilst the rest of the Commons half of MNIS may be firmly under the control of our crack team of librarians, preferred names are something of an outlier. Should a Commons member be lucky enough to have an honour conferred upon them, the Journal Office check if said Member wishes to include the honorific in their preferred name. Some choose to include, and some don’t. Which is fine. Unless, that is, you’re offstage in a government department or a committee office and you’re attempting to integrate data with a spot of string matching. Over the past few months, both of these sets of stakeholders have been in touch to complain that their spreadsheets were failing to string match when they assume ‘Sir’ and our Members do not. Following careful consideration with the Commmons Journal Office, it’s been decided to leave the current policy in place. After all, when you’re attempting to integrate data and find yourself string matching, something has gone wrong. A string being no substitute for a thing. As any fule know.
