Why Science != Science

7 comments

Ok, so I was thinking about this the other day and it seems that if a field needs to add the word "science" to its name, you can be pretty damn sure it aint a science.

Let's look at some examples:
Physics, Chemistry, Geology, Astronomy

Computer Science, Social Sciences, Political Science, Earth Science.

I don't think many would disagree that the four in the first row are sciences. But the second row? Well let me tell you, there's absolutely nothing scientific about writing a piece of code. And Earth Science!? They can't even come up with their own freaking theories! Every single one of "their" theories comes from Physics, Chemistry, and Geology.

All Earth "Scientists" do is memorize various physical phenomenon (i.e. glaciers) and the colours and texture of rocks. IMO, memorizing is not a science, unless of course you're doing a study on it in Psychology.

It's about time the official names of these "sciences" were changed from things like:
Earth Science
To
Earth "Science"

Peanut Gallery

>> Well let me tell you,

pk

>> Well let me tell you, there's absolutely nothing >> scientific about writing a piece of code.

There are two types of developers out there. Ones that use libraries to connect web pages to databases. And ones that write the libraries.

The first type are code monkeys.

The second type use the scientific process to choose the best algorithm. They try to make the interface extensible and less error-prone. They use a profiler to determine the performance characteristics of their implementation. They document and enforce pre and post conditions. They analyze exception safety. They use design patterns. They have their peers review their code.

A million trade school monkeys on a million keyboards can impliment Slashdot. But they won't be able to implement Posix, the STL, an OpenGL driver, Boost, the Apache runtime, the X protocol, etc.

>> The second type use the

bitch

>> The second type use the scientific process to choose the best algorithm.

Oh really, let me guess, you're the second type and you've also never connected a web page to a database. Well then, oh so mighty "second type", what is that "scientific process" you use to choose the best algorithm?

There isn't one. You use reason to determine which is the best one to use. You don't do experiments to determine the best one. You simply already know what algorithm can do what and you pick the best one for the job.

Oh, and by the way, of the "first type" and "second type" there are an equal percentage who do the following:

>> They try to make the interface extensible and less error-prone. They use a profiler to determine the performance characteristics of their implementation. They document and enforce pre and post conditions. They analyze exception safety. They use design patterns. They have their peers review their code.

And an equal percentage who don't.

Typical Code Monkey: Instead

sm

Typical Code Monkey:
Instead of writing a library, Monkey goes to apache or sf and finds a perfectly good library to use. Code Monkey writes a sweet app for KickAssedKompany in half the time as Type2 and gets paid 100 G's for his entrepreneurial ideas.

Typical Type2:
Writes a sweet peice of software, designing his own libraries along the way. Takes awesome software to KickAssedKompany who tell him they got a code monkey to write the same thing 10 years ago. Type2 goes home to hang out with his cat...

burn!

midge

burn!

>> You don't do experiments

PK

>> You don't do experiments to determine the best one.

It's called using a profiler and unit-testing.

>> It's called using a

ts

>> It's called using a profiler and unit-testing.

Dude, if you have to use a profiler everytime, then your findings aren't reproducible. And if they're not reproducible, then it's not much of a science...

"Archaeology is not a

ts

"Archaeology is not a science, certainly not an 'arcane science'. It's a discipline which employs (amongst other things) scientific techniques, such as C-14 dating."
Found here: http://science.slashdot.org/comments.pl?sid=110241&threshold=5&mode=thread&commentsort=0&op=Change

Is a perfect example of why CS is not a science...it's more of a dicipline.