2021 - Week 50

Our final notes of the year before putting aside our smart casual office wear and popping into party frocks for the ‘festive season’. Alhough, given current rumours around the ‘festive season’, quite frankly, who knows? We hope our dear reader makes the most of their time spent away from pixels. It’s always hard to know how to sign off on these things - “stay well” not having quite the same ring as “Merry Christmas”.

On the duplication of written statements

There was very little in the way of map making this week, which at least gave our eyes a bit of a rest. That said, Librarian Jayne and Michael added a few more bubbles and lines to our published draft procedure map and data - this time to capture ministerial statements announcing publication, made in the House of Lords. Of the seven such instruments we’ve seen to date, all seven have seen statements made in the Commons, and only one has also seen a statement in the Lords. We always said we’d duplicate where duplication happens, and so the Food and Feed Safety (Miscellaneous Amendments and Transitional Provisions) Regulations 2021 is now complete with a link to both the Commons and the Lords’ statement.

On the subject of statement duplication, Anya and Michael have dived back into the beginnings of our written statement model and neatened it a little. Much work is still required, the addition of a FRBR-style, work-level statement needing a very firm tyre kicking. We’ve added the work level statement because, with a finger firmly in the air, Librarian Emma suggests around 90% of written statements are duplicated across the two Houses. Having one work above two expressions would mean all the effort our crack team of librarians currently put into annotating would only need to happen once. It would also give us the means to know whether that 90% figure is true without having to stick fingers in the air. That said, it would require effort to create that grouping, so a little of what we might gain on the swings may well get lost on the roundabouts.

Remediating remedial orders

We have spent a considerable slice of the past two weeks staring in bewilderment at the legislation setting out the urgent procedure for remedial orders. It wasn’t at all clear whether such an order could be approved in the first 60 days, or whether the first 60 days were reserved for representations and the passing of an approval motion would have to wait until the second 60 days. I mean, try reading it. Does it seem clear to you?

Journal Office Eve assured us that the latter case was correct but urged us to contact Ellie and see what a lawyer said. So we did, and Ellie replied and confirmed Eve’s thinking. Ellie also kindly confirmed the new procedure labels for our beloved egg-timer and so - with some finicking and finessing from Anya - the egg-timer now comes complete with four different calculations for assorted flavours of remedial order. Marvellous.

Or at least, partly marvellous. In a separate conversation with Matthew, Michael asked for a drafter’s eye view of the same legislation. And well, agreement came in short supply. In Matthew’s view, if one understands that “statute is speaking”, then “the resolution can have been at any time in the 120 day period” and “the sunset and replacement are two different things, so it is perfectly possible to accept that the order is approved early but the Government still has to take representations and still has to replace the order if it considers that appropriate.” It feels like our New Year may well start with an email to Mr Greenberg.

Problems with parsing

If you tuned in last week, you’ll know we had something of a panic when our ternary logic-based parsing failed to work and nobody could remember why we’d taken a turn toward the ternary in the first place. This week, our Jianhan tweaked a couple of lines of code and we were back in the world of binary. Ish. Librarian Jayne proceeded to test and everything worked perfectly. Or more or less perfectly. Which was quite a relief.

This week, Jayne and Michael have been applying their procedural spanners and tightening the bolts on the bits that were slightly less than perfect. In short order:

We still need to preclude deferments of divisions from questions with established outcomes but there is only so much one can do in a week. So that will be our Christmas treat to ourselves.

In the interests of consistency and useful documentation, Michael has rewritten our Ruby procedure parsing code and turned it back into binary. Together with Jayne and young Robert, he’s also made a start on rewriting our design notes to reflect current understanding of how procedures get mapped and how maps get parsed.

The procedural and the political

An email chat with Journal Office Eve - again on the subject of the blasted urgent procedure for remedial orders - once more raised the issue that a thing being procedurally possible does not necessarily make it big-P Politically plausible. This is a problem we’ve chatted about at some length, but, to date, have done nothing to address. Whilst it is fine to parse a work package in the context of a procedure and say this step is now allowed to happen, in the cold, hard face of political reality there is every chance it is unlikely to happen.

Our efforts at parsing to date have been somewhat binary - a thing being caused to happen, allowed to happen or precluded from happening. For steps in the set of things allowed to happen, the hand of fate looms large over our light cone of the future possible.

To this end, Michael has begun to work on putting numbers against allowed steps to separate the probable from the possible from the preposterous. A new query has been written to get steps in a procedure and count the number of work packages subject to that procedure in which each step has been actualised. Dividing that by the total number of work packages subject to that procedure gives us a number between 0 and 100 indicating how likely a thing is to happen. 100 being probable, 0 being preposterous. It should probably be pointed out that this only deals with probability across the life cycle of a work package. For a work package ‘in flight’ it does not give us the probability of a thing happening in the imminent now. And, quite frankly, we doubt Michael has the maths skills to deal with this. We remain fairly confident we know people who do. Or might do.

It has since been pointed out by Anya, that such queries should only take account of concluded work packages, so the numbers aren’t biased by things in open work packages that couldn’t possibly have happened yet. So that looks to be Michael’s “festive season”. We hope yours is equally fun.