Rereading the classics

Last week I was a bit overloaded so this blog post comes a bit late, but better late than never. The last chapter of beautiful architecture titled "Rereading the classics" was a great read for me. The author were for the most part discussing object orientation and presented the programming language Smalltalk as an example of a system that had taken this to the extreme and made everything an object. I have never seen Smalltalk code before so it made for an interesting case study and I can definitely see how Smalltalk's legacy have shaped subsequent systems like Java and C#.

After having left the previous chapter, with its arguments for extending the type system through inheritance to solve most problems, with an uneasy feeling I was glad to see this chapter's careful use of it. I have indeed read Meyer's great "Effective C++" and have seen the penguins' failed attempts to fly before so I strongly agree that inheritance means is-a and is not a mechanism to compose code like the horrible COMPOSITE_PART of page 334. We learned to favor composition from our teacher this semester and the rest of the gang of four and there is no going back.

Smalltalk has a lot to teach us about Object Orientation and I find it very interesting, but not surprising, that much of the community that gave us design patterns and the guidelines for developing Object-Oriented systems that we use today came from the world of Smalltalk. I've mentioned this before, but things like this make me really wonder how much it is we that shape our tools and how much it is our tools that shape us. I certainly agree that language facilitate thought and surely this must mean that pure languages then facilitate pure thought.

The author closes with a too me surprising turnaround. After having talked up Smalltalk's strength and OO purity he makes the case for more pragmatic programming languages like C and C++ and even throws in a great quote by Stroustrup (although of course a Dane will prefer Kierkegaard!). He gives us some examples from the world of physical architecture where some of the most celebrated architectural masterpieces of the last century, like the Fallingwater house and the Villa Savoye, turned out to be completely unlivable for its inhabitants. Form must follow function and architecture must be made to facilitate the people who will live in it. It is like Alexander's villages in the Swiss alps where the naturally evolved patterns the buildings follow allows them to be beautiful while at the same time being functional for their inhabitants down to the last detail.

So is it for languages too and although I doubt Smalltalk developers will characterize it as unlivable neither has it had the same role on center stage as C++, Java and C#. However it, like beautiful functional programming languages with their purity and expressiveness, and T/16 and Erlang with their ubiquitous focus on reliability through replication it has served to inspire us and has influenced our other programming languages as well as the way that we use them. In some ways I believe the ideas that came out of the Smalltalk community have proved more powerful than any single programming language. At the risk of causing controversy I would like to draw an analogy to the socialistic democracies of Scandinavia and indeed to some degree most of present-day Europe. The ideas of Communism were very pure, but inherently flawed at their core. However, instead of implementing the anti-thesis of Capitalism, which is also very pure and although much preferable has its own problems, they landed at a hegelian synthesis of the two ideas, pragmatically borrowing what they perceived as the strengths of both with a market economy and a strong safety net. I do not endorse this one way or the other, but it is interesting how the radical pure ideas blaze for a while, but are then subsumed into more pragmatic approaches and becomes the status que that everyone takes for granted. I too, like Stroustrup, am inspired by the rationalists, but when everything is said and done prefer to live in a world based on pragmatic empiricism.