Gd idea - think of all the 'experienced' practitioners it wd put out of work, if anyone cd work it out for themselves with a gd bit of study!

Would it be 'black box' like PHPP, where you put in your proposed design and it rates its performance,
or wd it be instructional, setting out somewhat isolated first-principles, which you'd then have to combine economically and non-conflicting (come to think, that combining cd be what experienced designing is about)?

Your question reminds me of a book by the surgeon and writer Atul Gawande called 'The Checklist Manifesto' , about the use of checklists in the medical world. This quote on the subject says it better than I would:-

"New York Times’ Freakonomics Blog review
If there is one topic that I have no natural affinity for, it is checklists. I don’t use checklists. I’m not interested in checklists.

Yet, against all odds, I read Atul Gawande’s new book about checklists, The Checklist Manifesto in one sitting yesterday, which is an amazing tribute to the book that Gawande has crafted. Not only is the book loaded with fascinating stories, but it honestly changed the way I think about the world. It is the best book I’ve read in ages.

The book’s main point is simple: no matter how expert you may be, well-designed check lists can improve outcomes (even for Gawande’s own surgical team). The best-known use of checklists is by airplane pilots. Among the many interesting stories in the book is how this dedication to checklists arose among pilots."

A checklist isn't quite what you were suggesting but it might go some way towards it. Me no architect, but as an engineer there have been times when I would have done better to use one, e.g. 1. will it be too noisy? 2 will it be too heavy? 3. can it be maintained? 4. I like it - but will enough others? ......etc, etc. While it may not help with the assessment of the various factors, it could at least help ensure that none are overlooked.

There's one way that design differs from what an airline pilot or doctor does though - it tends to be an iterative process. So a checklist would have to be devised with this in mind. Certain things wouldn't be ticked off once but at each iteration. There'd have to be a way of preventing that from becoming too unwieldy.

Re Checklists:
It's not construction detailing but at the macro level pretty much all architects in the UK work to a common "Plan of Work" published by the RIBA:
https://www.architecture.com/knowledge-and-resources/resources-landing-page/riba-plan-of-work

Construction detailing was taught to us at the school of architecture I went to but not all of them do that. Over the last 30 years there's been a tendency toward increased specialism in design and now fewer architects than ever feel comfortable preparing complex detailed construction drawings. On larger projects these days big chunks of detailed design work are subbed out to specialist contractors or consultants.

Posted By: lineweightThere's one way that design differs from what an airline pilot or doctor does though - it tends to be an iterative process.

You will find that in these spheres, the checklists do (or have) change(d) iteratively. Incidents investigated by the AAIB for instance will often recommend (or demand) changes to checklists.

Posted By: borpin

Posted By: lineweightThere's one way that design differs from what an airline pilot or doctor does though - it tends to be an iterative process.

You will find that in these spheres, the checklists do (or have) change(d) iteratively. Incidents investigated by the AAIB for instance will often recommend (or demand) changes to checklists.

I don't mean iterative changes to the checklist but to the design itself.

When a plane takes off it does it once. You don't start with an outline taking off, refine it to a semi developed takeoff and eventually settle on a finalised takeoff (which the passengers then ask to be a takeoff from a different airport).

Posted By: gravelldE Myth Revisited

I was recommended that - "changed my life" - I thought it was quite loathesome! Employees as automata (unless 'revisited' means total revision of the 'original').

Posted By: lineweighta lot of what designing is about is weighing up priorities and 'combining' or integrating many potentially conflicting requirements including cultural and aesthetic ones which are not really quantifiable and would be outwith the scope of what I'm imagining

Bringing it back away from checklists, which are only one tool towards to 'combining' economically and non-conflicting.

I agree with "cultural and aesthetic ones which are not really quantifiable and would be outwith the scope of what I'm imagining" - I was def referring to technical 'combining'.

Posted By: fostertomI was recommended that - "changed my life" - I thought it was quite loathesome! Employees as automata (unless 'revisited' means total revision of the 'original').

That wasn't my takeaway. I think it relates to service businesses more than product ones (see the spin offs - E Myth Architect etc ).

Posted By: lineweightWould it, in principle or even in theory, be possible to approach construction details in a systematic way from first principles? That is, not relying on prior knowledge, experience, rules of thumb and so on.

In other words would it be possible to devise an exhaustive series of checks which could be applied to any detail to determine whether it will work? Even just taking one aspect - say for example, 'will it keep water out'?

Seems unlikely to me. There is too much variation in materials, situations, goals etc. Not to mention in the gap between as-designed and as-built.

There aren't many areas outside the hard sciences where the 'engineering' is complete and foolproof. Patients still die. Aircraft still crash. Buildings still burn. I'll bet that the breast screening scheduling software had the benefit of a raft of quality and project management standards applied to it but did they work?

Posted By: djhFirst principles won't help you decide whether the slope is great enough for a particular tile, or how many nails to use, or how much overlap between rows.

I would have thought that is exactly what first principles can tell you.
Then you add in the safety margins.
For any given tile overlap, known wind velocities, rainfall patterns you can work out how 'far up the crack' water can be drive.
Same for fixings, the mass of the tiles, its angles, the wind loadings, strength of nails etc should tell you what is needed.
That is what first principles is all about.

I think you're on a sticky wicket here, ST. (Exactly how sticky it is hard to say, using just First Principles). I can only see it being arguable if you upgrade Murphy's Law of The Cussedness of Inert Objects to the status of a First Principle - and a few other such laws along with it - and have precise knowledge of how these laws will work in any given situation.

If I'm to imagine a practical version of what I describe in the OP, then my definition of "first principles" could be fairly loose. I don't think everything would start from the laws of physics. To take the example of wind-driven rain, perhaps it would be a set of more general rules. Check for gaps of a certain type and orientation. Some guidance on minimum or maximum dimensions. Methods of sealing. I know this starts to sound like ending up back at rules of thumb or standard construction textbook, but what would be different is that there would be a methodical checking of everything. It wouldn't rely on you thinking to check something in a certain chapter of your textbook. It would also be focussed on general rules and an explanation of their purpose instead of example details to copy and modify. The danger of modifying an example detail is that you unwittingly modify something that you don't realise is critical to its functioning properly.

Also in a practical sense - there wouldn't really be a need to check established 'systems' like roof tiles. They are well understood. The critical points are where those systems join others. So wall to roof details and so on. This is where all the non 'standard' stuff happens because even if your roof tile system has excellent documentation with lots of example eaves detail it's likely that none of them will quite match exactly what you want to do.

One thing that comes to mind is leadwork. There are lots of very well established standard details for common situations. But sometimes I'll be detailing something that's nearly but not quite like one of those standard situations. Or maybe some kind of hybrid of two of them. In these cases what's correct is no longer definitive and you're left trying to use your judgement based on your best understanding of why certain things are a certain way in the standard details. But a more "first principles" approach might be better. The leadwork needs to extend X mm up the wall or down the roof slope *because* Y unless Z applies. Etc.

I found this book recently and realised it was everything I wish I'd discovered as a student: a plain speaking, first principles approach to details.

https://www.wiley.com/en-gb/Architectural+Detailing%3A+Function%2C+Constructibility%2C+Aesthetics%2C+3rd+Edition-p-9781118881996

Worth reading some of the excerpt on that link.

It's a little like the systems theory approach of Christopher Alexander in 'A Pattern Language', only applied to details. So that 'detail patterns' like 'Overhang and Drip' or 'Drain and Weep' can be applied to various different areas of the building.

Sounds like it might be what the original post was after, albeit not quite in checklist format.

Is Paisley a pattern language, a colourful Scottish.

Ah, thanks for that extra link, Tom. Chapter 1 certainly looks like good stuff, and seems to me to be pretty close to what Colin is looking for. I see Amazon and various other shops claim to have it in stock at a reduced price, so I'd be very tempted to buy a copy.