Im not sure why scientists are still publishing in Elsevier journals nowadays. Pre-internet, OK, but now I think journals should be following the model used by the Journal of Machine Learning Research, open on the internet.
In the end I'll have done the research, the writing, the editing and the formatting but of course the end product will in all likelihood be behind a paywall (this is sort of my own stupidity for rushing to get something published under the advice of a prof). There's no reason individual academic communities can't decide that they should establish there own reputable and open access journals.