I was working on ECG signals that were acquired in our institute and in a hospital to deal with all sorts of unbearable noises and artefacts. And then there were articles that would be published after testing their algorithms on a synthetic ECG (as if the world is lacking true hearts) with a cute gaussian noise (as if the world is lacking funky electrodes on sweaty hairy dudes), or the least noisy signals of the MIT database (the ones who test their stuff on the first seconds of 100.dat, you know who you are).

What I also found frustrating is that we have to trust the words of the authors about the results.

Cool, the signal is filtered but how long does your program take to deal with the 1.5 million samples? How many lines of code? What was the gear used?

These are all important things to know so that people know if they're improving stuff or not. But the FPGA cluster seldom gets mentioned in the articles.

I didn't expect I had to go through all that crap because I assumed they're frigging on renowed journals (so pre-filtered).

So I spoke to some people about it and it appears there are "international conferences" where you don't even have to attend for them to vouch you attended. You only need to pay.

After hanging out in some of the labs and seeing the mindset of the publishing game and how it's done, my supervisor still calls me so that we can publish something together but the answer is always that I don't have the necessary skills, nor the desire to publish.

I don't want to publish something that'll be a waste of time for other students.

There seem to be some efforts to deal with this racket (I think some companies are doing that)..