Faculty of Mathematics, University of Vienna
Professor of Mathematics, retired
During the fifties of the last century 80% of all mathematical journals were published by universities or learned societies; the rest was in the hand of commercial publishers. Large mathematical libraries contained most relevant literature; newly founded universities could afford to buy a complete mathematical library. At the end of the last century 80% of all journals were in the hand of commercial publishers; some of them are the most profitable enterprises worldwide. As a result, all university libraries have financial problems and none has complete runs of all relevant journals anymore: the journal crisis is upon us.
The open access movement tries to move financing on scientific journals from libraries to authors or their funding agencies as a way to solve the journal crisis; at the first glance this looks good.
But the leading commercial publishers have quickly turned this into another income stream without real added value for the journals. Many new publishing houses started to scam scientists with emails asking for papers to be published in new open access journals ("only 3 Dollars per page").
I receive now several such emails per day; the "Antarctica Journal of Mathematics" is one that I remember. Moreover, publishing under the open access model is somewhat under the shade of "buying a publication".
A solution of the journal crisis: Libraries of major mathematical institutions could publish and finance one journal each which is freely accessible online. In my view, an optimal organization of scientific publications could work as follows: authors first upload their papers to a major freely accessible preprint server like arXiv.org. Editors and their trusted referees look there for papers which meet their interest and report on them. The editors then invite the authors to publish in their journal and the author chooses a journal where he is invited. Major impact papers could then be invited to be republished in "Collections of significant papers".
All my own scientific papers are freely accessible via my personal homepage. For that I have scanned all pre-digital era papers. All digitally born papers are also available via arXiv.org. Since 1990, in each book contract I have asked that the work becomes freely accessible online after 4 years or 1200 copies sold. As my scientific estate, my homepage will go to our library and to emis.org who will be asked to keep them accessible for eternity: these will be my electronically collected works.
In doing this I follow one of the recommendations of the Committee for Electronic Information and Communication of the International Mathematical Union: http://math.umn.edu/~olver/ceic.html.
Looking for a result or a paper I first turn to MathSciNet or Zentralblatt-Math. If my University library has access, I can directly click for the pdf. If not, I look into arXiv.org or look for the personal homepage of one of the authors (an 18th century approach). Scholar.google.com is also helpful, but it offers books only page by page; one needs clever researching.