January 26. 2009 13:10

First, that great guy C.A. and his ideas on patterns, derived from applied engineering, design and construction of housing - why would IT reference this here and there, at all? Why not Plank or Goedel or whoever else? What's wrong with us? Are we so great that we can build any product that can be compared to old castles, cities with great architecture, or something similar? Why should we think that ideas of Alexander are applicable to IT? I doubt that IT achievements can be put on the same level as achievements of architecture.

Then, Darwinism is not applied correctly in your posting. Theory suggests that three mechanisms drive evolution - natural selection, genetic drift, gene flow. There's zero genetic drift in IT life (obviously), and strongly limited gene flow, since even in vertical descendance in the same product line of the same company not always can be found what we call PLA - Product Line Architecture. Left with natural selection, we have idea that not only changes that enables species/products to adopt to conditions of the environment are left to live, but in favour are changes that give specific competitive advantage in given environment against other beasts, and enhance reproduction. So natural selection, would it be applicable to IT, should be promoting the designs, architectures, products, which have traits resulting in better "living" in our human-owned IT universe. Fortunately or not (depending on your preference for philosophy of life), in this universe there's no mother-nature which is careless of individual species, have equal rules for all players and favours only fact of life per se. What we have is predefined goal of IT products existence, where "living" is targeted to bring money, or publicity, or respect/etc to their creators; and we have the target audience which decides - to live or to not - for each product, based on generally very clear preference - better to have working piece of crap today, than diamond-shining masterpiece tomorrow.
So I don't see how clarity or usability or better understanding is promoted by such kind of natural selection that we have in IT. In fact, ugly, messy, and unusable things will go longer distances just because they do solve the task - deliver the value to customer today, and generate revenue for the owner (forgive me father IBM, but most of your software offsprings are falling into this category).

However the idea of appealing artefacts with clarity and usability as built-in gene is quite attractive. Even more, the clearer the architecture of some platform, the more it spreads to minds of developers, and the more it enables to build on top of it, the better we would label it. And big IT vendors establish their own evolutionary processes, supporting them via mass-production solutions, which generate revenue and attract developers to their niches. And the more developers support the architecture or platform, the more it will reproduce, so here it is similar to natural selection, indeed. What's funny though is that in IT we even have a kind of religions, philosophic-like community thinking which tries to explain the existence of specific development and why it is "righter" than competition.

Pheeeww. Did I just manage to write a comment larger than initial posting, dammit?

Andrew

January 26. 2009 16:35

A lot of things depend on how much money was invested into marketing: it's not enough to create an appealing stuff, you need people to know about it as well!

z00m

January 29. 2009 16:03

Andrew,

Thanks for the write up! Lots of cool ideas popped up so here go my comments.

There's zero genetic drift in IT life (obviously), and strongly limited gene flow

Think about memes: units that transmit from one mind to another. I believe memes play the same role for our minds that genes play for our bodies. Memes can replicate, mutate and spread. The artifact that "tells a better story" spreads better. Though if the story isn't inherently coherent enough, the artifact (or the story) will mutate on the way.

Think about misused development architectural and patterns. Weren't they misused because their memes mutated? ;)

Next, I absolutely agree with you about a very clear preference - better to have working piece of crap today, than diamond-shining masterpiece tomorrow. But think about evolution again. It's slow. It painstakingly mutates and filters out the genes (memes). Working piece of crap will never, ever, live through its time.

Of course, there are exceptions to this rule - but if you look at them closely, they aren't exceptions actually. They probably are "pieces of crap" wrapped into appealing artifacts, that takes away the smell.

In fact, ugly, messy, and unusable things will go longer distances just because they do solve the task

Hopefully I partly answered above. The only thing why it's possible - because the tasks get changed before the evolutions kicks in! However, my post was about the developer/architect view, not the producer/customer view. In the dev/arch world it's hard to sell an "ugly, messy, and unusable thing": evolution is much faster there.

Andrew

January 29. 2009 16:12

z00m,

Unfortunately you're right - to some extent. Another idea is that remarkable things get noticed over time. But it doesnt mean you should wait for recognition after you've done some Remarkable Thing.

Creating appealing stuiff over and over again - that's what matters and eventually pays off.

(I hope )

Andrew