Twenty-odd years ago I held subscriptions to perhaps a dozen ACM and IEEE publications. (SIGPLAN, SIGARCH, SIGOPS, SIGGRAPH, SIGCHI, IEEE Software, Communications of the ACM, and an irregular Pascal/Modula-2 publication come to mind … there were probably others.) Those publications were at that time, the only source of high-quality high-volume information about computers and software.
I canceled my subscriptions in the very early 1990’s. At the time it seemed as though the articles were starting to repeat, and offered mainly predictable observations. Put differently, I was no longer in “catch-up” mode (my area of study in college was Physics, not Computer Science). There were still things to learn, but not through the journals.
My ACM and IEEE memberships expired a long time ago.
Today we have a flood of information on the internet, not all of which is of high quality, but if you know how to look, the quality is there.
Lately I have run across a few references that lead to inaccessible documents owned by ACM, IEEE, and other similar organizations. The article was inaccessible without paying a relatively steep fee. In no case have I had any reason to believe the article referenced was worth the price asked.
Once publishing in such journals meant rather a lot. Peer-review meant the quality of the articles was generally high. Individual and library subscriptions meant your work and ideas got the widest possible (relevant) distribution. Storage in libraries meant the greatest availability for the longest time. There was no better alternative.
Today those journals are - for the most part - a ghetto. Isolated from the vast stream of knowledge and discussion on the web, restricted access journals are very nearly irrelevant. What sane person would subscribe to such a publication?
For academics, the publication in such a journal was (and is, presumably) the almost the only way to gain status. Once, this was a good thing. Much of the most interesting work was done in or near academia. Professional journals offered the best path for disseminating that knowledge out to the broadest body of practicing professionals.
This is no longer the case. There is a wealth of direct sharing between professionals on the web. Very few are likely to subscribe to the old journals (published in print - how quaint), which means the journals are largely academia talking to itself. The old model of restricted access journals is clearly wrong.
There is a place for peer-reviewed publication of material archived so that long-term reference is possible. But the material needs to be out in the open, on the internet, to be in any real way relevant. Not sure what the economic model would be to support such an archive, though the long-term costs for storage and access are becoming less relevant. Perhaps such an archive should be replicated by universities as a matter of course (much as universities are expected to have substantial libraries). Come to think of it, to be most relevant, such an archive should on occasion use hindsight to acquire submissions. A work of value found “in the wild” should be archived along with all works of equal significance.
In any case, it seems that the old “publish or perish” dictate is far less worthwhile if the work is only seen by other academics.