Tumbling off the peak of a truly amazing week, this one saw our wave crash into the more usual troughs of map making drudgery. Not that we complain. And certainly never whine. Drudgery, after all is more in tune with the times. That said, week 5 did have its upsides as our social lives expanded and we actually spoke to people.
Before we start, we should establish that by social life we mean talking about work related matters to a different set of people. What’s work and what’s life seems less established than it once did. And the boundaries of work-life balance inevitably blur if one combines red wine at midnight with talking to ex-colleagues about parsing work package maps.
On Wednesday evening Librarian Anya, young Robert and Michael were lucky enough to borrow the brains of Samu (ex of this parish). As you might guess, the talk revolved around procedures, logic steps, work packages and the parsing of maps thereof. Robert and Michael are now on their third iteration of pseudocode. Or fourth. It’s hard to tell. Anya has thrown a slight spanner in the works with a requirement to separate steps that cannot yet happen - a question being put on a motion that has not yet been tabled, and steps that cannot now happen - EVEL certification in light of suspended standing orders. This means we need to taint all the roads on unreachable islands, to let the computers know they are currently untraversable. This was explained at some considerable length in last week’s notes and we have no intention of typing all that nonsense again. The new requirement lead Michael to rewrite our handling of business steps to take account of the state of the inbound route. After chatting with Samu we now think our approach might best be described as, ‘brute-force breadth-first traversal’. It certainly does not seem particularly efficient. Samu, who has the advantage of knowing about maths and such things, suggested we might instead look at depth first traversal. If we do go down that road, we’ll need to think some more about when we tell the machines to stop traversing a cyclic graph. And quite frankly, we’ve been too tired this week. Way too tired.
Additional red wine being poured, Samu and Robert headed off to walk the Seven Bridges of Königsberg, a problem not eased by two of the bridges being bombed in the war. It was agreed that the problem we face is not best solved by Michael’s rather clumsy attempts to place loops inside loops. What we actually need is a dab of Prolog and a touch of Constraint Logic Programming. Michael was much cheered to hear Robert announce his intention to learn Prolog, although his confidence evaporated somewhat when Robert added, “again”. In the meantime, we have material to read and ponderings to ponder. Thanks Samu.
Nursing only the mildest of hangovers, the next day saw the same threesome and a bright-eyed Librarian Jayne meet with TNA colleagues Catherine, Matthew, and John. Or ‘Samu’s new boss’ as we like to call him. The meeting had originally been planned to talk about if and when TNA might be able to supply machine readable procedural duties to help us determine the appropriate map to pick, and to feed our egg timer with all the data it might need to calculate scrutiny periods. Not yet being the answer. But things have moved on since we first proposed the meeting, and we had new questions. Such as: are proposed negative statutory instruments just another manifestation of any statutory instrument that might follow? And how might we bypass our powers and duties model to point directly at enabling Acts? On the latter, a new sketch was sketched and Librarians Anya and Jayne and computational experts Robert and Michael have a new task to add the shortcut to our legislation model. On the former, everyone present seemed to agree that our current approach of two work packageable things having two layings and generating two work packages subject to two procedures, should in fact have been one work packageable thing with two layings, two work packages and two procedures. But John did suggest we chat to Luke about this. Luke - having written chapter, verse and book - might be of help to us here. Luke has very kindly offered to give us a beginner’s guide to statutory interpretation and a promise he’ll try to help with our more specific problems. Thanks TNA folks. And thanks Luke.
Well rested and fresh as daisies, Friday brought a meeting with Andrew to chat all things peerages and how we might interlink concepts in Parliament’s Thesaurus with items in Wikidata. Our main problem here is the separation of the peerage from the peer. Our second problem is gender. Wikidata tends to be person centric. Our concepts tend to represent the peerage and not the peer who happens to be holding it. Except in the case of hereditary peerages being initially held by females, or initially held by a male but with a special remainder allowing them to be held by females in the future. In which case, a new concept is added to the thesaurus to reflect the change in rank label as the gender changes. Wikidata has records for around 750 of the current Lords’ membership, modelled as people holding a position. Which we think is around 50 short of quorum. It also has some items for the more abstract notion of a peerage, particularly where Wikipedia describes such things. How we make links between Wikidata’s peers and peerages and our - occasionally gender split - peerages will require more investigation into how Wikidata handles peerages that switch gender and a bit more thought. Lots more thought.
If you’ve been following along from home, you’ll probably be aware that David was kind enough to give us access to the peerage data that sits behind his website. Michael has been poking at said data - partly to better understand what our Lords’ membership model should look like, and partly because it’s fun to poke at. He’s written a series of rake tasks to reshape the data from fairly flat tables conflating peerages and peers to a new schema that picks apart the two and pops a ‘peerage holding’ in the middle. This week he’s been mainly battling with the notion of subsidiary titles, which exist in David’s data as child records of peerages. In the main this works fine, but breaks for a chap called Scarsdale who held two peerages, one of which has a special remainder. This allowed the two peerages to part company somewhere along the inheritance chain. David’s website seems to suggest that the thing that groups a ‘parent’ peerage with its subsidiary titles is a single letters patent. So Michael has proceded on this basis and grouped accordingly. This works well for the vast majority of cases, as subsidiary titles have the same date of letters patent as their ‘parent’ peerage. Except … we have a single outlier in the shape of the current Duke of Cambridge whose main peerage has a letters patent date of 26/05/2011, and two subsidiary titles - the Earl of Strathearn and Lord Carrickfergus - which have a letters patent date of 29/04/2011. Whether this is a data entry error or a misunderstanding on our part remains unclear. Anya has checked the London Gazette and found a single announcement. So for now, Michael has assumed incorrect dates on the subsidiaries and smushed them into the same letters patent as the ‘parent’ peerage. If you’re one of the very few people who understands how letters patent work for subsidiary peerages and what might have happened with the Duke of Cambridge, please do get in touch.
Michael has also bumped up against the gender switching peerage problem. David’s database handles the first holder of a peerage only. But we want the option of adding subsequent holders of a peerage. And, if we’re to cope with a peerage potentially switching genders, the fact of gendered ranks in a peerage title is problematic. For hereditary peerages initially held by a female, Michael has gone with female rank / male rank title. For peerages initially held by a male but with a special remainder allowing for female inheritance, he’s gone with male rank / female rank title. It’s possible that there is established practice on this and that Michael is not following it. If that’s the case, he can only plead that he can’t find it.
Standing orders finding meme fame, we can’t deny it was quite tempting to start with this section this week. But in all honesty, we don’t have much to say on the subject. Tom, from the ParlRules project, was kind enough to read through and idiot check the comments on our standing order model. No corrections were issued. Thanks Tom. If you, dear reader, are a standing order kind of person - preferably not representing Handforth Parish Council - please do take a minute to read them. READ THEM AND UNDERSTAND THEM.
Jayne and Michael have continued on their quest to logicify our procedure maps. This week it was time to turn their attention to the draft negative statutory instrument map. Made all the easier because it’s basically the made negative with the word ‘being’ switched out for ‘becoming’, and a new top and a new tail. The logicified map is now published for your delectation. It still needs labels for non-business steps, and we still need to shove it all inside the machines. Rewarding tasks, eagerly anticipated.
Wednesday saw a mass meeting of librarians with Anya, Jayne, Claire, Emma and Ayesha joining computational experts Robert and Michael for a chat about factors that might cause the scrutiny clock to stop. Other than time itself. Currently, should an instrument be revoked, we reset the ‘clock end’ step to the date of revocation. But we do not do that for other things that cause a procedure to conclude, such as the approval of a fatal prayer, the rejection of an approval motion, or the withdrawal of the instrument. Given we already have ‘procedure concluded’ steps, leaving the clock to describe time parallel to procedural possibilities feels like the more sensible decision. We still need to confirm this with JO Jane. Emails will follow as emails inevitably do.
Hot off the heels of last week’s chat with JO Jane, we’re keen to implement start steps, concluding steps, key steps and Member action steps at the earliest possible opportunity. But the more properties we add between procedures and steps, the noisier the model becomes. Robert and Michael hit upon the idea of turning four predicates into a new class with a type enumeration. But we already have a class to cover the grouping of steps into stages in a bill procedure. Which is not dissimilar to grouping any other type of step in any other type of procedure. So that model has been repurposed as a step collection class with a step collection type class to cover the enumeration. And everything is much neater.
Team Anya is in the process of expansion. If you’re an organised and organising kind of person, looking for gainful employment and you’ve made it this far through these notes, please do consider applying. The opportunity to work with us is not to be sniffed at.