Live data from Hacker News

Christopher Alexander has died

cnu.org

121–130 of 139 posts

Re: Christopher Alexander has died

#121
post #9
post #3

Massive PDF of A Pattern Language: https://arl.human.cornell.edu/linked%20docs/Alexander_A_Patt... But I recommend owning and reading it in print. Great book to dip in and out of at whim.

I don't remember which volume it was from, but doing any general maintenance I always remember this suggestion of his: Work on the most neglected area first -- e.g., work on the most neglected outdoor area of a property first. I don't recall his specific reasoning, but in my mind the essence was: Otherwise you will be acting on displacement anxiety and When an eyesore is taken care of first, your perspective on poten…

Maybe pattern 104, "site repair"?

> If we always build on that part of the land which is most healthy, we can be virtually certain that a great deal of the land will always be less than healthy. If we want the land to be healthy all over - all of it - then we must do the opposite. We must treat every new act of building as an opportunity to mend some rent in the existing cloth; each act of building gives us the chance to make one of the ugliest and least healthy parts of the environment more healthy - as for those parts which are already healthy and beautiful - they of course need no attention. And in fact, we must discipline ourselves most strictly to leave them alone,so that our energy actually goes to the places which need it. This is the principle of site repair.

Re: Christopher Alexander has died

#122
post #36

Earlier quoted context omitted.

He was where the concept of Design Patterns came from. In the early days of Design Patterns (GoF, et al ), he was often quoted. I purchased a couple of his books: A Pattern Language [0], and The Timeless Way of Building [1]. These books were very readable. [0] https://en.wikipedia.org/wiki/A_Pattern_Language [1] https://en.wikipedia.org/wiki/The_Timeless_Way_of_Building

IMO Alexander's design pattern concept is intelligently superior to the terms trivialization in the atrociously highly rated GoF book. GoF enumerates the tedious patches one needs to use on cumbersome languages like Java or C++ due to the languages primitive nature. The main advantage of GoF that now when you say "Factory pattern" everyone knows sort of concept they are speaking of. However. In the scope of Alexander…

You immediately see the difference when you open A Pattern Language and the very first pattern is:

> Wherever possible, work toward the evolution of independent regions in the worId; each with a population between 2 and 10 million; each with its own natural and geographic boundaries; each with its own economy; each one autonomous and self-governing; each with a seat in a world government, without the intervening power of larger states or countries.

No software pattern work has the audacity to start by suggesting that the ordering of the world's governance is relevant to the creation of good software.

Re: Christopher Alexander has died

#123
post #32

Earlier quoted context omitted.

OOPSLA talks of that era were fantastic. Alan Kay has often paid tribute to Christopher Alexander in his talks and I cannot resist mentioning this from Alan from the following year https://www.youtube.com/watch?v=oKg1hTOQXoY

Too bad Kay has never paid enough tribute to Nygaard and Dahl in the same manner. His comments from that '97 OOPSLA talk on Simula being incomprehensible are absurdly misleading.

will investigate. any and all pointers welcome.

Re: Christopher Alexander has died

#124
I read about Alexander's Norwegian barns and their "quality without name" and how that perhaps should apply to software too. And surely it applies, but in practice is not practiced very much.

One reason this might be is that software applications are more complicated things than barns. And the kind of complexity that matters here is not really in the implementation but in the interface. To design great applications I believe you have to consider that user is just one part of the "system". It is not only software running on a computer, it is also software running in user's head and fingers and the protocols users can and will use to interact with the rest of the system.

When you design a software application you must decide how much and what will the system do and how much and what will the user do. That is a complicated decision and it is not easy to come up with the optimal choice. Possibilities are almost endless.

We don't usually consider the user as part of "software" but really they are, they are just parts in a bigger system with formal ways the user-part can and will interact with the computer-part of the system. Interaction with a barn is not trivial but quite simple compared to computer-based applications, I would think. But I agree the quality without a name is important to strive towards. It is, or would be, when everything seems to perfectly work together in different contexts and for different purposes.

Re: Christopher Alexander has died

#125

Every time I meet an architect or designer, I ask them about Alexander, and in the sample of about 12, only one (a world renowned architect) had heard of him. He's hugely influential in software, and was a big influence on me in the 90's and 00's, along with Tufte, Hofstadter, and Feynman. I think it was hacker culture that really bouyed his ideas and kept them in the popular consciousness, and in turn, his ideas bec…

I learned about his work through reading The Not So Big House. I've also asked Architects and Builders I've met if they are familiar with his work or the related Not So Big House movement, I haven't met a single one that have heard of either. It seems so strange to me. It would be like meeting a programmer who has never heard of functional programming. Even if you choose to use OOP instead of functional programming, surely you've at least heard of functional programming. How can you be an architect and be ignorant of Christopher Alexander??

Re: Christopher Alexander has died

#127
post #5

Saw him speak at Berkeley around the time his book on the Eshin project was released. Talked in very strong terms about the importance of working in a holistic, caring way. A SWE in the audience spoke up, saying basically "look, we want to follow that approach, but it's hard, we have a lot of stakeholders to satisfy in order for a project to happen." Alexander was unyielding. "Once you've worked with love, you won't…

This is the reason I changed how I approached software projects. Alexander made me rethink how software should be built: with love for the user, but also with love for other developers.

Adjust your own mask before helping others with theirs.

It always appalls me when developers reason that a software development system that barely functions for themselves will somehow result in an application that the users will enjoy paying to use.

Software that doesn’t work for the developers will eventually stop working for everyone else too.

Re: Christopher Alexander has died

#128
He's lovely but: He writes the bible on good design but makes the insights proprietary.

None of his works have a Creative Commons license. All his works are "all rights reserved" as in:

"let there be light, but you mustn't quote me"

Fingers crossed: his daughters let Alexander's good thoughts flow free...

Re: Christopher Alexander has died

#129
post #9

Earlier quoted context omitted.

I don't remember which volume it was from, but doing any general maintenance I always remember this suggestion of his: Work on the most neglected area first -- e.g., work on the most neglected outdoor area of a property first. I don't recall his specific reasoning, but in my mind the essence was: Otherwise you will be acting on displacement anxiety and When an eyesore is taken care of first, your perspective on poten…

Maybe pattern 104, "site repair"? > If we always build on that part of the land which is most healthy, we can be virtually certain that a great deal of the land will always be less than healthy. If we want the land to be healthy all over - all of it - then we must do the opposite. We must treat every new act of building as an opportunity to mend some rent in the existing cloth; each act of building gives us the chanc…

Yes, that is almost certainly the origin and what I was thinking of -- a site and land context originally.

As a heuristic (extending possibly to so-called technical debt?) I believe it generalizes.

I am lately learning the delight of another pattern: Outdoor Workshop.

Re: Christopher Alexander has died

#130
post #123

Earlier quoted context omitted.

Too bad Kay has never paid enough tribute to Nygaard and Dahl in the same manner. His comments from that '97 OOPSLA talk on Simula being incomprehensible are absurdly misleading.

will investigate. any and all pointers welcome.

Simula67 was already known in academic circles by the early 70s. Donald Knuth had first taken an interest in Simula I after a visit in Oslo but couldn't manage to bring it to Stanford because of the very high licensing fees the Norwegian Computing Center were charging for it. That first iteration of the language was more focused on being a general purpose "system description" language for different kinds of real-life simulations (this would also kickstart Nygaards work on the social impact of technology and the beginnings of user-oriented system development, then participatory design). Simula67 however wasn't the "incomprehensible" language that Kay describes. The second iteration was much more focused and had all the elements of modern object-oriented patterns: classes, subclasses and inheritance, objects, object references and attributes, object dot notation, polymorphism, etc.

So Kay's take on Simula being this strange language he had to "make sense of" for his work on Smalltalk, or even the fact he claims he coined the term "object-oriented" is total BS. Here's a more nuanced and impartial account of Simula by James Gosling, if you want to know more: https://www.youtube.com/watch?v=ccRtIdlTqlU.

Post reply on HN