2024 - Week 12

Using hypertext (to subvert hierarchies)

Being big fans of prog-rock behemoths Gnod, we occasionally kick off our latest release with a brief reprise of the final track on our last album. Last week, we concluded with not so much of a humblebrag as an actual brag - that these very weeknotes had been described as a ‘work of genius’ by the bloke that wrote the Agile Comms book. Thanks Giles.

Imagine our delight then, when Anya and Michael finally put down weeknote pens, only to find we’d made another appearance in Giles’ output, this time under the heading ‘Examples of good weeknotes’, where, we’re told, our dear reader leaves these pages both ‘entertained and informed’.

This is not the BAFTAs. We’re not here to bask in our own glory. Or, perhaps, not only for that. This week, we’ve been thinking about why we write weeknotes. We think Giles is partly right. Being fond - indeed protective - of our many and varied stakeholders, we’d hate to think of them having to read Yet Another RAG Report. It being hard to believe that anyone has ever knowingly asked for a RAG report. Writing in a longer form - and yes, we do go on a bit - does seem to have caught the attention of more than one regular reader. JO Jane often gets in touch about something in these notes, putting us right on minor and major bits of procedural pedantry. We hope she stays in touch on the other side of her impending retirement. Correspondents include such luminaries as Journal Office Eve, Journal Office Matt, Legislation Office Liam and Researcher Richard. The next time you get welcome clarification on business processes because of a RAG Report, do get in touch and we can trade stories.

It is though more than this. Organisations tend to be far more porous than they ever admit. As Michael likes to say, when he was at the BBC, the Backstage project was just beginning. It was mostly a technology thing, but somehow it turned some bits of the cathedral into something roughly resembling a bazaar.

Whilst we’re handing out advice, always remember, in any organisation there are more experts out there than in here. So it’s not just internal stakeholders we hear from, but people who’ve now left Parliament, like Arabella, David and Paul. And people who were never in Parliament like Ben, Steve and John - we warmly recall the time John wrote to say we’d given him nightmares. All of them have helped us along the way and we thank them.

We also write weeknotes because we’re of the old fashioned belief that the people who should get the credit are the people who’ve done the work. Far too often, project documents reduce workers to job titles on some Gantt chart or its more modern equivalent. We try not to do that in these parts.

Finally, we write weeknotes because we enjoy writing, because we’re proud of the work, because we think the work is interesting and because we think other people might find it interesting too. No matter how small the audience.

If we were to add to Giles’ tips and tricks for weeknotes:

Librarians of the Week

This week, our much coveted Librarian of the Week trophy goes to Librarians Anna and Emily, whose excellent librarianship saved the psephology ship from sinking. Anna took Monday off, only to return to a torrent of emails from a slightly panicked Michael, asking if it might be possible to apply MNIS party IDs across six ginormous spreadsheets covering three general elections. By the end of the week.

Our crack librarians clocked on before Michael had had chance to consume his second coffee and tenth cigarette of the day, and they didn’t clock off until Michael was safely ensconced in the Two Chairmen. A magnificent effort quite deserving of our super shiny trophy. Not only had they added MNIS IDs for all parties everywhere - a requirement for plugging in notional results at some point in the not too distant - they also tidied up both party names and party abbreviations. A side-effect of which means we’re now running with a single Libertarian Party, where we once somehow had two.

They also completed work on our Electoral Commission to MNIS party ID mapping spreadsheet. A spreadsheet we’ll need when the time comes for Data Scientist Louie to export candidacies from Democracy Club into our own database. A spreadsheet without which Parliament’s general election planning and the future of representative democracy may well have been in peril. So that’s a relief.

Working with librarians is never not a pleasure, and yet even we are astounded when they pull out all the stops. Or ‘put in the hard yards’, as Young Robert would no doubt put it. All of which brings us neatly on to our continuing psephological explorations.

People, places, parties

Spreadsheets tidied and annotated, Michael once more poached some pub WiFI and got stuck back into his beloved database and import scripts. A quick rewrite later and the importer was MNIS party ID compliant - at which point, the moment of truth arrived. Dear reader, we are happy to report that everything loaded beautifully. Which means the data modelling, data wrangling and data import for the 2015, 2017 and 2019 general elections is now considered complete. At least until we notice the next thing.

In terms of visible changes, Statistician Carl kindly donated a party colour spreadsheet - all handily mapped to Electoral Commission IDs - meaning our party listings have now taken on a rather pleasing Power, Corruption and Lies vibe. Or at least that’s how it looks if you’re roughly the same age as Young Robert and Michael.

Carl was also kind enough to point out that Michael’s attempts at D3 visualisations were, in fact, terrible. Even the one Michael quite liked. We wouldn’t wish to get Michael started on the subject of his ‘colour-divergent’ eyes again, so we’ll note only that he had no hope of knowing this. Over a pint in a pub, Data Scientist Louie and Michael concluded that the best course of action would be to strip out the D3. At some point fairly soon, Louie intends to break out the Shiny and make new visualisations. This time taking a rather more professional approach. Once they’re ready, we will include them on the website. Or ‘transclude’ them, as Young Robert would probably say.

Still with statistician Carl, a wee while back he asked for result listings by English regions and countries. All quite simple. Setting aside the definition of a country - really, let’s not go there - Carl also asked for aggregations at the level of Great Britain. Which added some complications to endeavours. We’re happy to report that that work is now done and we have all the usual listings - including by parties - under both countries and English regions.

Another small change, this one requested by boss-boss-brarian Bryn, means our Parliament pages now link to the boundary sets in operation at the time. Which is a quite a nice request and one we were more than happy to comply with.

With the to do list getting ever shorter, attention has turned to how the thing looks, which - depending on time and effort between us typing this and you reading it - will be either awful or awesome. We remain firmly of the belief that one day, HTML brutalism will return to fashion. After all, those 1950s concrete blocks seem to attract all kinds of middle class fanboys. The Rococo flourishes of the moderne web page says nothing to us about our lives. “Was Audrey Hepburn famed for her makeup?” we ask. No, she was not. Still, beauty is obviously in the eye of the beholder, and not everyone favours bare-metal markup.

To that end, boss-boss-brarian Bryn leant a helping hand with a dab of Parliament-themed CSS. Michael tried to apply Bryn’s efforts to our code base, but the last time Michael touched CSS he’d just started at the BBC and been tasked with rebranding JazzFest 95 to JazzFest 96. JazzFest 95 being red and JazzFest 96 being green. Or the other way round. Not a job well suited to the ‘colour divergent’. Young Robert has now taken over CSS duties, which feels like a much better solution for all concerned. May all your pixels be polished and all your divs turn semantic.

Finally - being as old as we are - we’ve also added robots.txt. Sadly it would seem, not an item you find on the lists of most modern project managers. Or product managers. Or delivery managers. Whilst we were on that job, we also added humans.txt, because, if we’re about anything, we’re about giving credit where credit is due.

Slightly to the left of psephology, but related at some level, Librarian Phil asked whether we should treat candidates lacking a description on their statement of persons nominated as independents. A quick chat with Louie and a scan through what Democracy Club do confirmed his intuition and that is what we now plan to do. Splendid.

Meanwhile, Librarian Emily has been off researching ‘higher-level geographies’. Higher, in this case, referring to geographies above a parliamentary constituency but below an ITL 1 region. What’s currently in MNIS is a proper hodgepodge of counties and upper tier local authorities and lower tier local authorities and local authorities that are both upper and lower tier at the same time. Before we can tidy that, we need to first understand it. Happily, Emily now has a pretty good grip. By this time next week, she’ll be more than capable of holding an intelligent conversation with resident geographic expert Neil. Given the hideous complications of UK geography, that is not to be sniffed at.

Trello card ping-pong between Librarians Jayne and Ned and developer Jon, continues at a frankly frantic pace. Freshly flung onto the done pile this week, we have Church of England Measures, deposited papers, the quaintly named paper petitions, parliamentary proceedings, formal proceedings, research briefings, written statements and European scrutiny ministerial correspondence (House of Commons). Excellent progress all round.

Over in new old search backend land, we’re still attempting to cope with fallout from our Solr upgrade. As we’ve noted on many an occasion, the code as installed when that project closed in 2016 has since been, well, utterly abandoned. Which means we’re attempting to upgrade across six major versions in one leap. It can’t come as a surprise if we occasionally fall on our faces. Given we appear to be in the mood for handing out advice this week, here’s some more: if your supplier upgrades their software, upgrade when they do. If you don’t, you’re only gonna end up staring into the rear view mirror, tossing expletives at your past self.

On the subject of which, those of you who have met Young Robert will know that he taught himself to swear through close reading of the books of Enid Blyton. Which is why Room 13 is currently echoing with cries of ‘FIDDLESTICKS!’ and ‘BLASTED THING!’ as he twiddles the knobs and dials of Solr 9 in an attempt to make it not consider apostrophes an abomination. Still, it makes a change from our crack team of librarians, most of whom swear like sailors.

In slightly better news, our RSpec tests are now considered complete, having gained a new method to compare text snippets across old Solr and new Solr. Ned and Michael attempted to get it all running on Ned’s machine - his personal machine, not his corporate Dell, obviously, that would be ridiculous - but that has, so far, proved impossible. More work for next week.

Panic stations: on written corrections, ministerial and otherwise

With Hansard about to start publishing written corrections from Members-who-happen-not-to-be-Ministers on the other side of the Easter Recess, poor Jianhan has once more found himself knee deep in rusting pipes and murky puddles. That the ingest into Parliamentary Search is one more example of code that hasn’t been touched in 13 years isn’t really helping matters here.

Being a diligent chap, our Jianhan is clearly concerned about deploying patched-up old code to even older servers, and suggested we upgrade the boxes before we do anything else. Clipboard in hand, Delivery Manager Lydia headed back to the Change Advisory Board with not one, but two, Requests For Change. Requests For Change that were, happily, acceded to. Another blocker unblocked. Let’s see what the next one is.

It is, one supposes, a little like having the plumber round to put in new taps, only to return home and find your bathroom ripped out. No doubt an experience we’ve all had. So far, Jianhan has resisted the temptation to whistle through his teeth and ask, ‘what cowboy put that in?’. We suspect that moment will come. It is at this point that our knowledge of plumbing ends and our tortured metaphor runs dry. We put our trust in Jianhan and our team of librarian testers (thanks Librarians Anna, Claire, Jayne, Ned, Phil and Steve). After all, what other choice is there?

Subject specialists and how one finds one

The House of Commons Library is not some room with some books. It does have some rooms with quite a few books, and it is, primarily, a research library. It exists because, whilst Members who serve in Government have access to the ‘might’ of the Civil Service, Members who do not serve in Government do not. Across the Library there are subject experts in all kinds of areas, from transport to housing to defence to what counts as our constitution. How then, might Members and their staff find the right specialist for the right subject? Enter the Subject Specialists Directory, a short pamphlet akin to the Yellow pages but with more Library researchers and far fewer plumbers.

Our regular reader will know that our crack team of librarians maintain our beloved Parliament Thesauri, a tool used for subject indexing parliamentary material for the purposes of search. And, rather sadly, not used for subject indexing any of the other things that get subject indexed. Specialists and enquiries being just two examples. We are then on a mission, with Librarian Susannah leading the charge, to liberate the taxonomy from its software constraints, allowing it to be used in other systems far and wide.

Late last year, family friend Silver and his Data Language colleagues created a small demo - taking our taxonomy, a list of researchers and their Data Graphs product, stitching it all together with a dash of transitivity and building a Subject Specialists Directory that ‘meets the expectations of the internet age’. As Young Robert might say. As Young Robert might also say, that work was a ‘safe to fail probe’, exploring how we might ‘liberate and leverage’ the taxonomy. ‘Sweating our assets,’ as Young Robert might add. He reads a lot of Gartner, does Young Robert. Pragmatic chap. Talks a lot of business sense.

We’re pleased to say the demo worked, but more on the basis of ‘leverage’ than ‘liberation’. The dataset we took was static with no plumbing in place to keep things in sync. So now we’re onto stage two, putting the pipes in place to liberate the data. It will probably come as no surprise that our taxonomy management system is another piece of software that hasn’t been touched by human hand in 13 years. So we’re attempting to upgrade that whilst also plugging it into other systems. Or ‘fixing the plane in flight’ as Young Robert might say.

On orders being standing

If you’re one of the very many people waiting on tenterhooks for our persistently citable standing orders to appear, we can only apologise. At some point, we ran clean out of ways to pay Shedcode James. As in, we had the money, but no legal means of procurement. Luckily, consummate unblocker Librarian Susannah joined the team and our economic embarrassment has now been cleared up. All of which means, James is back on board, fit and working again.

So long was the interval between James doing the initial work and our ability to pay him to do more, by the time we sat back down with the editing side of the application we were baffled. It appeared to work perfectly. It did everything we expected it to do. But, somehow, it was damned difficult to use. Chief standing orders data wrangler, Librarian Claire, convened a pixel-based meeting with James, inviting computational playboys, Young Robert and Michael, along for the ride. All eventually became clear. Our data model is a tree structure: fragment versions belong to one order version, order versions belong to one revision set section, revision set sections belong to one revision set, revision sets belong to one House and business extent. All quite simple, relatively speaking. But at some point, we appear to have added links traversing more than one step in the tree. So trunk to twig, as it were. Whilst these links may have saved on clicking time, they utterly confused our mental model of the domain. James is now in the process of amending the editing side of the application, removing any and all shortcut links and reinforcing the tree-like structure through the addition of breadcrumb trails. It all goes to show that ‘minimising clicks’ is never the goal and that domain semantics live in the hyper as well as the text.

Unfortunately, because many of the changes are inside the admin interface, there’s not much we can really show you, so you’ll have to take our word on improvements. A couple of changes are publicly visible. The first being the removal of fragment counts from revision set titles. No matter how much of a standing order obsessive one might be - here’s looking at you Martyn - exposing the number of fragments in a revision set definitely felt like too much information. James has also cleaned up our order and fragment pages, which are now both leaner and cleaner. Top work Claire. Top work James.

At this point, we should probably point out that, yes, we are currently running with test edit data. Please do not read too closely. Also, please do not make any procedural decisions based on this. Mouses play no formal role in Parliament and we believe this is unlikely to change.

I am a procedural cartographer - to the tune of the Palace Brothers

With Librarian Jayne busy elsewhere, we don’t have an awful lot of cartographic news this week. For which our dear reader may be breathing a sigh of relief. Or possibly not. Maybe they’re actually thinking, “I used to really enjoy weeknotes, back when they went on and on about applying ternary logic gates to parliamentary procedure. They’ve really gone off the boil.” Well boo sucks to you, dear reader. We can only work with the material we’re given.

Still, as procedure adapts, we’re always willing to adapt with it. A couple of weeks back, Librarian Ayesha spotted that the Joint Committee on Human Rights had reported correspondence with the Minister of State for Prisons and Probation on the UK-Philippines Transfer Agreement. This was - at least to the best of our knowledge - the first time the JCHR had taken this step. And we were there to take it with them. A new business step has been added to our CRaG treaty map, awaiting actualisation the moment the correspondence is published. Stay tuned guys.

Jayne being Jayne, she has, of course, added a new item to our list of committee correspondence SPARQL queries. Would you expect anything less?

British people in Zoom chatrooms

Readers who are lucky enough to have these notes delivered direct to their inbox may have noticed a recent change to the subject line. What was something about procedure modelling is now titled ‘fables from a Cynefin Stable’. This is because we’re big fans of Dave Snowden and big fans of everything Dave Snowden says and does. Did you know he can smell snow? Of late, thanks to repeated prods from Dan, we’ve also been giving some thought to training a language model on matters procedural, getting as far as writing a reading list and not much further. Imagine then our excitement when news came through of an ML, AI and ethics workshop courtesy of the Cynefin Company.

Late on Wednesday afternoon, Anya, Ben, Robert, Silver and Michael settled back in armchairs, fresh brews to hand, ready to listen and learn. The session began with what the host chose to call ‘provocations’. A word Young Robert hasn’t heard since he found himself part of a student sit-in Paris, back in ‘68. Second of the provocateurs to appear on stage was Dave himself. We use ‘stage’ loosely, Dave being propped up in bed in a guest house in the Lake District. Kudos, Dave. Kudos.

Things were going quite well, until the host announced we were all to be sent to ‘breakout rooms’ to discuss the ‘provocations’. With people we’d never met. Our local Slack instance lit up with barely concealed horror. Or, actual terror. Don’t they know we’re British, we said. Or at least from the British Isles. We can’t sit round discussing ‘provocations’ with people we’ve never met. It’s embarrassing.

So one by one, we left, unable to confirm if the ‘provocations’ provoked anything beyond terminal shyness.