When you did industrial research, how many things involved just one paper? I'm not trying to create an argument. I just accept that academic research is not as pure as many people portray it to be. I used to think the university/research system was a pristine place of people doing research for the sake of research. I don't think that anymore.

I'm not by any stretch saying that academic research is pure, at all. But I don't see how that is relevant.

It feels akin to my surprise in other threads where some people insist e.g. that there is inherently something "bad" when a company like MS releases source to the linux community, because they didn't do it purely for giving things away and there must be some ulterior motive.

Why does it matter? I guess it matters to people if they perceive the world as a 0-sum game, so there is always a "take" side to the "give".

But I don't see it proven that the world indeed functions as a 0-sum game. We are all scratching our own itches, we perhaps all have ulterior motives in "giving things away" and yet we can earn a living while the world at large benefits from the fact that we did give things away.

For example:

Some time ago I had to add code to the linux kernel because of problems in autodetecting the volume resolution of a particular audio codec IC. Either no one had discovered it before or it was not a priority for other people, since there was no fix for it at the time. But it was a problem for my client, so they paid to fix it. There was no hesitancy in open sourcing the patch, since it is a building block and them giving it up has 0 implication on their business.

Also they had an ulterior motive in doing so, because having the fix in mainline meant they would not have to keep merging it in while staying up to date with kernels, and would not have to worry if certain code structures change at some future point, etc.

Since they had run into the problem it is likely that someone else would also have run into it at some point; and by that patch being available, someone else is going to be saved some time and expense (except for the $0.00012 or whatever "distribution" fee to download the code), and maybe that person gets to spend an extra hour with his kid.

Who loses in this case?

Or more recently where a client had a problem with an SoC where there is a timing problem in the macrocell relating to the interrupt controller, and boards based on that SoC would mysteriously stop working under some conditions. Again this was a problem, they paid to get a software fix for it, and the fix will be submitted to the community. Again there is an ulterior motive, because more people will get to test & debug the fix on a wider variety of environments which will save the client time in the long run; and again this is going to save headaches, hassles, money and possibly lawsuits for other parties.

So in both cases an "impure" "giving away for free but with ulterior motives" and everyone stands to gain?

Obviously academia is of a different scale, but it seems to me more of the same thing -- hey we need to publish in order to improve our career prospects, but how does anyone loose from this?

And to answer your question: it varied. On one short project I purchased six papers (iirc); on one multi-year project I purchased one. In the latter case much of the research had already been embodied into available standards, and often times it was possible to pop an email right to the author.

Many years ago I was doing research in TCP over satellite connections in Africa; shot an email to Mark Allman in the USA, and received their full experimental TCP Vegas implementation by return email. Who looses here?

I just don't know how else to phrase this; we all keep going on with our work but there are many things we can actually share (that may directly benefit us and appear as ulterior motives); that ends up making life easier.

Surely this is way more efficient than the alternative (not sharing), even if you have the "middleman/distribution channel" taking a nickle?