Hmmm. I'm very critical of much academic publishing. At its worst it is lazy, egotistical, vanity publishing. At its best it's usually inadequate.
Pretending that most of what is known, is what is "published" in the correct "refereed journals", is a self serving myth that masks and undervalues so much of what is really known. In ICT in particular the papers are often so far off the mark as to be risible; papers don't simply lag behind exceptional, but effective, practice, they miss it altogether. I sit on enough editorial boards to be allowed to be depressed about this.
Asked by a media interview for my "best ten things in education in the last decade" I was not surprised to realise that 8 of them hadn't been written up in "proper" papers.
This would be alarming if reputation was all that was at stake, but too often funding also depends on this myth that the lazy academic writer is the most worthy of support so that they can write even more about what they might, but so rarely actually will, do.
On the other hand a few papers really do shed light and genuinely introduce new voabulary, new uinderstandings or retrospective reflections of real activity, real research and real success. I wish mine were always that good (!) but in this category of the RetroBlog I'll be parking just a representative sample of academic outputs that I think mark a route through some real progress in new learning. I've tried to pick interesting and readable ones too, from both journals and book "collections".
And just in passing , why is it when really big action research projects reach really sound, demonstrable,, unequivocal conclusions, that people ignore them anyway?... Probably because you don't get published by offering agreement, you get published for "new".