“A pattern language” explained (2016) | Hacker News Reader

“A pattern language” explained (2016)

permaculture.co.uk

I have been an admirer of Alexander's A Pattern Language book on architecture for 40 years.

Even then, this is new to me. I recommend you read this article, even though it's deep and difficult. It can be read as a fundamental criticism of UI/UX as it's practiced today. In particular it could be seen as a rejection of Jonathan Ive's (Apple's) design principles, principles which I have increasingly found disturbing and lacking in human relevance.

An interesting question would be: after 40+ years, what happened? Has the book positively impacted how we build homes, neighborhoods, cities? (I'd say very little)

His main theme is: build for humans (can apply to anything, software also).

But today we are all about status. We think what we possess is who we are. This is a lie created by marketing. But it's very hard to change this in our cultures.

We now see movements like the tiny house movement, minimalists, healthy food movement and so on.

Imho those are all anty riactions to the marketing we are now hearing for years.

So even if Christopher is right about how to build it can be very hard to change the way we think (about building).

Alexander himself agrees with you. The pattern language approach in his hands and in the hands of some architects was a tool that produced wonderful buildings. In many hands, the patterns were there and the buildings were still miserable. The patterns were necessary but not sufficient. I posted a comment elsewhere in the thread referring to his recent work, 'The Nature of Order', where he thinks, after forty years of work, that he's figured out the sufficient part as well.

Here is an impressive review of a huge number of scholarly critiques of Christopher Alexander's work: https://cityterritoryarchitecture.springeropen.com/articles/...

Personally, I've never been able to take him seriously enough to work up the patience to analyse his work to anything like that level of depth. His project is hubristic.

I got half way through that article, and everything up to that point had been strawmen, so I gave up.

As for hubris, what is hubristic about trying to figure out how to build well?

I used Alexander's book to design (and build) 3 homes. I feel they were exceptionally nice homes.

There’s a book specifically applying the “pattern pattern” to UX, namely Jennifer Tidwell’s Designing Interfaces (O’Reilly)

It’s 10 years old or so by now, but has held up better than most technical books of its age.

I think it went out in 2012, so that makes it 6 years old. The field has developped enormously since then but the principles probably hold quite well.

I agree that self-assembly, auto (as in self) generation is what's lacking. Not sure what form it would take in software, but I feel that all the seams in man-made objects are limiting progress in tangible forms large and small. I suppose in software, these boundaries are apis and protocols, which ought to be stitched together in a more organic, self-learning way.

We see very different things. I'll have to re-read the article from your perspective.

Would be nice to have a meta-protocol to negotiate protocols. Probably looks something like game-theory.

Just finished watching a great 1996 lecture by Christopher Alexander: https://www.youtube.com/watch?v=98LdFA-_zfA

I recommend it for those who have read his work, but never seen or heard the man in action.

If this resonates at all with you, go get Alexander's four volume work 'The Nature of Order.' And stick with it. It's a lot of ground to cover. The goal, as always for him, is to be able to construct built environments with a timeless, life enhancing quality, and he spends volume one demonstrating that the quality is repeatable among individual observers and then constructing a system of analysis to let understand what drives repeatable observation.

Then volume two is about how you go about constructing environments with this quality. If you have a hard time, flip ahead to the sequence for constructing a Japanese tea house. Volume three is how you do this in practice, and how you set up the logistics and project management so it is possible. Volume four tries to construct a mindset which lets the approaches in volume three emerge as obvious behaviors.

Each is a step back: here is this desirable, repeatable quality; here is how you generate a state with this quality; here is how you generate systems that generate states with this quality; here is how you generate societies that generate such systems.

It's utterly brilliant. And I want to warn off everyone planning to translate it immediately to software. There are sequences of unfolding that we can develop and use (I've got some notes I've been working on), but the quality that we are trying to generate is not the same. The timeless, life supportingness that Alexander analyzes is driven utterly by Euclidean geometry. Software architecture is not. There is a quality that we are looking for and that, much more about melding invisibly into the domain and being adaptable as the domain adapts. I really think that the quality we're looking for is buried in relations among Wittgenstein's language games, not in geometry.

I also think we would benefit from looking a little further afield than just constructive sequences. He talks about the Tahitian canoe maker's chant which gives the order of unfolding of making a canoe. Let's compare that to the Epic of Lata, used by the sailors of the windward Solomon islands as their sailing instructions and protocol when sailing the alien looking Vaka o Lata (see http://vaka.org/). When they are embarked, they take the roles of characters in the epic and reenact the parts of it that correspond to what's going on. In the wonderful book 'How Buildings Learn' (if you like Alexander, you'll love that one), there are vernaculars documented of how traditional houses of various kinds expand in repeatable ways over their lifespan. Similarly, we spend more time adapting software and operating software than we do building software from scratch, so our unfoldings will need to be much more shaped by a vernacular driving how a system gets reshaped.

Mind you, I may just be sleepy and that rambling may be totally inconsequential and useless.

Wow, thanks for a very insightful comment. Curious: are you an architect? Or, are you into software perhaps?

My father is a structural engineer and master carpenter who designed several houses, and I worked with him as a kid. Then twenty something (cough mumble) years of software for me, with digressions along the way.

"Thriving Systems Theory and Metaphor-Driven Modeling" (see https://www.springer.com/gp/book/9781849963015 ) is a book that attempts to bring the ideas from the Nature of Order in the world of software.

Interesting. I'll have to read it and see if gets somewhere useful.

>the sailors of the windward Solomon islands as their sailing instructions

That bit reminded me of a great book, Morris West's The Navigator. It's both a good adventure story (Robinson Crusoe style, but with a motley crowd) and probably a parable or lessons for people.

A Pattern Language is required reading. Ever since reading it I've dreamed of making some kind of rule-based generative architecture program- where a building is created according to selected patterns.

Hmmmm. Interesting.

Reminds me of meta-algebra. For example, a^n - b^n = (a - b)(a^(n-1) + (a^(n-2))ab + ... b^(n-1)) .

Reply on news.ycombinator.com