The coffee pail bubbles down to an acrid stew. The prawn sandwiches are curling. Night pulls the sun under the horizon. And the meeting, the interminable meeting, goes on. Civil servants clip in and out with scraps of paper. Hangdog faces dominate the table. Increasingly impolite apologies are offered, as the protagonists nip into the corridor to field hissing mobile phone calls, anxious not to be overheard. Nerves fray as weariness increases. Everything heads south. The final text is assembled, scrappily, bit by bit. Disagreements are finally squared, or not squared. Minds soon turn to the aftermath, to the lines to take, and the victories and the strategic retreats to be spun.
This, give or take, must be the atmosphere in the Smith Commission as the last few grains of sand fleet through the hourglass, and Saint Andrew's Day approaches. The political representatives are due to announce their heads of agreement by the 30th of November: just days away. I'd expect this document to be simple, categorical - and leave several points of detail and practicality underdeveloped. It will not, I imagine, be a detained administrative blueprint, but a broad statement of principle and intent. How could it be otherwise, given the breakneck speed with which the report will have been assembled?
The Commission hasn't exactly been leaky - but there have been a few noises off from the political parties participating in it, who must now have a fair idea about the nature of the proposals which the group is likely to endorse, and the key points of friction between them. One of the weekend's most interesting stories was the news that the Liberal Democrats have reverse ferreted to an undisclosed extent on welfare devolution. So much, so vague. The new commitment may be much less than it seems, and until Thursday, one for the Kremlinologists trying to work out which coalitions of interest have formed within the Commission - and the extent to which Labour or the SNP finds their respective maximalist and minimalist visions of autonomy isolated by the other participants.
It now seems probable, for example, that income tax will be devolved in its entirety - but important questions remain outstanding about the autonomy Holyrood would enjoy in terms of setting the basic rate and the bands. Elections too, seem likely to be devolved, with some hints that the fixtures and fittings of the Scotland Act will also be liberalised, to allow the Scottish Parliament to bend and shape its structure to meet changing needs.
Pensions, however, surely have a snowball's chance of finding their way onto Holyrood's balance sheet (and in the absence of significant revenue powers, this may well be a damn good thing too). I'd also expect to see heads of agreement on codifying the Sewel convention. I'd be surprised if the legality of any future independence referendums was clearly entrenched in the new Scotland Bill - but you never know. There will be other flotsam and jetsam too, of greater and lesser degrees of technicality and public interest. Minimum wage? Answer came there none. Abortion law? According to the Sunday papers, maybe.
But the political firecracker is social security. What precisely does "substantial new welfare powers" look like? What can it look like, in a historically unitary system where social security has been squeezed within the framework of a universal credit? I'm reasonably confident that the Commission will agree in principle to something along the lines of the additional general welfare competency or a top-up function which I commended to them in my submission earlier this month.
The idea is doable, legally workable, good in principle, and in line with fundamental principles of devolved autonomy and "pooling and sharing" resources. We might gripe - not unreasonably - that such an authority would be a paper responsibility only, while public spending is in the vice. But in a small way, it would beef up the armoury of the Scottish Parliament. Would it represent a "substantial" new power? Well, yes and no. In the absence of new control over significant swathes core welfare spending, it would be a drop in the ocean. But not without utility or significance for all that.
There will be a strong expectation, at least, that the much-mooted housing benefit and attendance allowance will be devolved. The bedroom tax has given the scheme a high political salience, and expectations have been raised. But we can't be too blithe about its devolution. It'll cost a fair whack to wheedle housing benefit out of the universal credit - who will bear the cost of those changes? And if the cost is prohibitively high, and would consume an unreasonably high percentage of the Scottish budget alone, would it be worth the bother?
It is one of those unlucky conjunctions for the Union. Dissatisfaction with welfare reform has increased the political salience of social security policy, and demands for its devolution, while Iain Duncan Smith's consolidation of welfare spending into a universal credit has made the devolution of these powers considerable trickier and more costly. If only Calman had had a little more forward vision back in the late 2000s. In terms of core welfare spending, the Smith Commission is now in all or nothing territory, essentially. And that has implications for the tax base and tax powers too. There is no point investing Holyrood with responsibilities it can't afford on a sustainable basis.
Given the Labour Party's case for the Union, its "pooling and sharing" patter, however, it is difficult to see how any devo deal which includes Iain Gray and Gregg McClymont can be particularly radical, falling closer to everything than nothing on key forms of social security for the unemployed and the disabled. If Smith can't cajole the UK parties into a bolder vision, welfare policy looks doomed to become the Union's running sore. If the Union remains fragile, a feeble Smith Commission offer will do little to staunch the flow.
The structure I propose - a general grant of power to Holyrood combined with protecting the UK welfare enactments under Schedule 4 - also represents a constructive way of approaching other issues, such as equalities, empowering Holyrood while protecting core UK statutes. Equal opportunities are currently reserved in Schedule 5. A better way of proceeding would be to protect the Equality Act 2010 from modification or repeal by Holyrood, but to give the Scottish Parliament the green light to adopt regulations and prohibitions to promote its own conception of the demands of equality.
As things stand, the Scottish Parliament is tied up in pointless prohibitions by the Scotland Act's Schedule 5 approach. The time has come to conceive of the relationships between Holyrood and Westminster more flexibly and imaginatively. It chimes with the home rule methodology that autonomy should be devolved, save where there are compelling reasons to reserve the power exclusively to Westminster.
There are several areas of policy, in a Union, where compelling reasons can be found. Bones should not be made about that. The "devo-max" vision of devolving everything short of foreign affairs and defence was always over-egged. What about our common market, the heart of the 1707 union, and everything that comes with it? What about decision-making on currency?
The radicalism of the Smith Commission deal will depend, to a very significant extent, on which side - the autonomy-promoting Nationalist contingent - or the reluctant devo-minusers of the Labour Party - the other Smith Commission participants coalesce around. The Nationalists can be expected to argue that the settlement - any settlement - falls short of the "master key" of independence. Quel surprise. You'd expect nothing less.
But there are other propaganda wars which the Unionist parties must fight and win. For the Unionist spinners, the plans cannot be another badly-sold Scotland Act 2012 - finnicky, grudging, technical and forgotten. It must be crisp, easy to understand, its real-life relevance clear and categorical. It must be bullet-pointable.
The proposals must be seen to honour the ambitions of "the Vow," however vague its commitments and mercurial its terminology. The bare terms of "home rule" and "federalism" - though much-quoted - don't really take us very far, compatible with a range of more and less powerful devolved institutions. Indeed, Holyrood already enjoys greater autonomy in some areas that many federal institutions in other states.
But politically, Labour, the Liberals and the Tories must be seen to be as good as their word. The SNP can be expected to slate the proposals as falling short of independence. The party would struggle, however, to make the broader and more significant indictment - that the Vow has been ratted on - if Smith's proposals are and appear to be substantial and bold. Smith seems likely to be the end of road for devolution: an event, not a process. Ruth Davidson's line in the sand, if you like. Thus far and no further.
As the delegates sit down for their final meeting, slurping the cold coffee and labouring towards dawn, welfare autonomy perhaps represents Lord Smith's greatest challenge. Room for manoeuvre curtailed by the centralising force of the universal credit reforms, bounced by Labour's historic opposition to devolving key, highly visible benefits, it is increasingly difficult to see how the Commission's recommendations on Thursday can avoid falling short of the expectations raised.
Cry last orders, ladies and gentlemen. The Commission is in for a long night on Wednesday.