The Gate

ChristopherAlexander said "Once we have built the gate, we can pass through it to the practice of the timeless way". To ChristopherAlexander, his pattern language was supposed to be TheGate to the timeless way of building. He did not think that his pattern language WAS the timeless way of building, but he hoped that if people followed it, they would be led to the timeless way of building.

What makes you think we can ever get through that gate? -- PeterMerel

The timeless way is what (some) people really do. His pattern language is supposed to help other people learn to practice the timeless way. Now, perhaps you could argue that there is no timeless way, that the qualities that Alexander appreciated existed only in his own mind. But if you agree that they exist, and that some people are able to make buildings with those qualities, then the natural question is "How can I do it?" or "How can we teach other people to do it?" ChristopherAlexander thought his pattern language would teach people how to do it. He did not think that his pattern language completely captured everything about the timelesss way of building, but he hoped it captured enough.

The proof that his pattern language is the gate to the timeless way of building would be if people studied it and then were able to practice the timeless way. ChristopherAlexander did not think that this happened. Thus, he would probably argue that his pattern language was NOT the gate to the timeless way of building after all. Nevertheless, it was a reasonable goal at the time.

Any methodology attempts to be a gate into some area of expertise. "Read this book and you will become a better object-oriented designer!" Obviously some people become experts, so something must be a gate. However, a lot of people seem to be unhappy when you say that TheGate is to practice your craft, hang around with experts and have them critique your work, and after enough years, if you work hard enough and are smart enough, you will become an expert yourself. We all want a silver bullet.

-- RalphJohnson

"The Timeless Way" is more KungFuTse than LaoTse. The very first line of Lao is (more or less literally)

 TheFlow? that flows is not some timeless/cosmic/abstract flow
This is thought by many to be a direct reaction against confucian notions of ideals and perfect orders.

So perhaps Alexander is making a basic confusion in thinking of a TimelessWay?. I hasten to add that's not to say his ideas aren't good, useful and fascinating. But perhaps we're not doing them justice by thinking of them as "timeless". Scientists don't pretend their theories are "timeless". They just think they're the best description they have right now.

So perhaps we ought not to regard any pattern, even alexandrian patterns, as timeless, or producing timelessness, but more that they are just our best practice right now. A new technique, say, might render any of them obsolete tomorrow. Just as it took a long time for people to think of a WimpInterface when all their previous engineering was geared to CommandLineInterfaces.

I think this idea is in the Alexandrian notion of a PatternLanguage; that a PatternLanguage is a self-consistent map, but that any two PatternLanguages might not be commensurable. -- PeterMerel

ChristopherAlexander didn't claim his book was timeless. He claimed that there was a way of making buildings that was timeless, that could be observed in buildings thousands of years old, and he was trying to explain that way. His pattern language was an attempt. He thought that particular patterns were part of the timeless way of building, but he did not claim that his description of them was.

Although lots of people seem to think that Alexander was influenced by various oriental philosophies, I do not think it is necessary to look to them to understand his writings. If it helps you, fine. But don't claim he is confused because he disagrees with one or more of them. He is probably using words in a different way than they are. -- RalphJohnson

According to JimCoplien on ZenConcepts, Alexander openly admits the relationship to these oriental philosophies. But as to timelessness itself, you seem to have ignored my phrase, "or producing timelessness". More than that, I don't understand why you're focusing on Alexander's descriptions of patterns - I haven't suggested anything about Alexander's books or descriptions in this regard. -- PeterMerel

Then I have no idea what you are talking about. I have been writing about Alexander's ideas, so I assumed your questions were about them.

Above you say, "ChristopherAlexander didn't claim his book was timeless. He claimed that there was a way of making buildings that was timeless [...] He thought that particular patterns were part of the timeless way of building, but he did not claim that his description of them was."

I am talking about Alexander too, please don't misunderstand, but I don't see why you draw this distinction between his text, rather than his ideas, and his pattern descriptions, rather than his patterns. I didn't draw that distinction in what I wrote so I don't understand why you're addressing it. I'm sure it's not a StrawMan, so I figure I ought to ask what you're driving at with it. -- PeterMerel

Patterns are concrete in nature and can represent a best practice typically presented as a solution to a problem in a context. But some patterns are more timeless than others. For example, when I think of the CommandPattern, I feel that it will be more timeless because programmers will continue to find a context where they want to decouple a request from its invoker. However, when I think about patterns for distributed computing, I see that many of them rely upon the fact that you are aware of location, persistence specifics, and whether what you want to use just vanished. When the programmer is abstracted from these issues because of advances in both hardware and software, they may not be as applicable. -- PhilipEskelin

I agree that CommandPattern seems pretty timeless. ObserverPattern and many other DesignPatterns too. I have no way, however, of knowing that this is so; folk are constantly surprised by obsolescence. Once upon a time, FlyingButtresses seemed timeless, but you see precious few of them on modern skyscrapers. Why should we think that some new technology (say, QuantumComputing) won't obsolete Commands and Observers too? I guess I'm wondering what benefit we get from thinking of timelessness, and what benefit we might get from thinking of timeliness. -- PeterMerel

I feel to make CommandPattern that timeless, you would need to make it more abstract. As a pattern, which has concrete applicability, it would become less effective. That's like saying you should abstract FlyingButresses? so the pattern lasts forever. Perhaps it's a difference in context that adjusts a pattern's applicability, and not the tendency to make something more abstract that is inherently concrete? -- PhilipEskelin

I think you're onto something here, though I have no idea how to compare the gauge of contexts. Maybe by the size of the PatternLanguage if each of its pattern descriptions is as abstract as it can be and still cover the context. But no more abstract than that. OhOh?, where have I heard something like that before? -- PeterMerel

View edit of December 27, 2004 or FindPage with title or text search