6. mars 2005

3/4 useless crap 1/8 total bullshit but 1/8 genuinely useful

Today I finally got convinced that the majority of all the numerous "research" papers out there are just frecking useless crap. Just an unimaginative collection of random words, arranged cleverly in the most obscure and useless way possible.

I actually think that some good "researchers" unconsciously or not write papers in such a way that it is not humanly possible to understand a word of what they are trying to say. Sadly sometimes these researchers are really smart people and their research interesting but they are so afraid that their colleagues will suddenly think that if they write their results in an clear and understandable way that they and their work is stupid or dumb.

So vast amount of people write hard-if-not-impossible to understand self-masturbation literature, filled with useless and obscure mathematical equations. Why? Stop it!

I have during this weekend and most of last week been struggling to write a literature review of a common machine learning method called "Boosting". It is not a complex technique and I found a lot of papers on the subject and actually ended up reading about 7 of them, and by reading I don't mean skimming. I began with papers titled "A brief introduction..." and "A short tutorial..." but they where all filled with fancy obscure wording and overly complex equations (btw if you use the sigma operator in an equations, write a short description following it) that made the material they where trying to introduce incomprehensible. Late last night and this morning I had almost lost hope of finding any useful material on the subject.

It wasn't until this afternoon that I finally stumbled onto a 60 page paper that I had previously overlooked. Finally! a well structured paper explaining the technology in a clear and concise manor without all the fancy notation. All the theoretical proofs where cleanly placed in their own chapters and only the absolutely necessary equations where presented outside of them (and always a good explanation followed). I read the whole thing in about 5 hrs and learned more than I had in the previous 5 days. Shit! wasted all this time and for what? Nothing?

Why is this necessary? When will people realize that writing a GOOD and useful research paper requires more knowledge and work than writing a bad one.

Before the next deadline, really think about it...

Engin ummæli: