This post originated from an RSS feed registered with Python Buzz by Jarno Virtanen. Original Post: Computing is not doomed, I think Feed Title: Python owns us Feed URL: http://sedoparking.com/search/registrar.php?domain=®istrar=sedopark Feed Description: A weblog about Python from the view point of Jarno Virtanen. Latest Python Buzz Posts Latest Python Buzz Posts by Jarno Virtanen Latest Posts From Python owns us

I am beginning to form a rule for texts that declare that the current state of computing, or software development, is doomed and something completely new is needed in place. The rule goes like this: Whenever a text declares that the current state of computing is doomed, I will stop reading the text.

You see, I do not think that the current state of computing is doomed and I do think that every security flaw found in some Microsoft product is not a further proof that they are doing everything wrong. Not because these claims are definitely not true, but because either of which is not a falsifiable statement. Sure, many software projects are not on budget nor time, and there are security flaws in Microsoft's products, but I claim that these cannot be attributed to anything particular, at least not without a proper comparative analysis.

It seems that the reason is thrown away as soon as a wider perspective is taken. Typically, like this:

• 65% of software projects are late, source [Xyz84]
• 80% of software projects are over budget, source [Zyx86]
• where did we go wrong?

And then some anecdote:

• typical example is company Foo & Bar
• spent 20 million dollars
• over-budget, over-time, flawed product
• many other similar cases

And then:

[a cute quote from some distinguished scientist; probably from Thomas Kuhn]

• we need a fresh start
• I have thought about this a lot
• my approach attacks these problems

I don't buy it.

There is no evidence that the same, ie. projects over budget and over time, wouldn't be the case with the supposedly new approach. Even a case study is better than nothing, although they might be a bit flawed, like the one comparing Haskell, Ada, C++, etc in software prototyping productivity by Paul Hudak et al. (Interestingly enough, Hudak's homepage says that "the paper was rejected from a journal because the experiment is full of all kinds of holes".)

Science advances in increments and its increments should be verified by sound scientific experimenting. And no, software development per se is not science, but observing its effiency and characteristics should be.

(Sure, Kuhn argued that once in a while science will advance through a paradigm-shift, a revolution if you may, and the supporters of the old paradigm will not understand the new paradigm through normal reasoning and argumentation. But I don't think such has been the case with these people who say that the current state of computing is doomed. But I might be wrong.)