In fact, many times after visiting Project Euclid, I saw DOI after visiting the paper and clicking on "about".

I am not sure whether this was changed recently.

And example where this works is here: projecteuclid.org/journals/duke-mathematical-journal/volume-95/…

Sometimes there is no DOI. I thought that this was for publications that do not have DOI.

Here is an example when Project Euclid doesn't list a DOI: projecteuclid.org/journals/…

> Freeman J. Dyson. "Missed opportunities." Bull. Amer. Math. Soc. 78 (5) 635 - 652, September 1972.

I provides a MR link: mathscinet.ams.org/mathscinet-getitem?mr=MR0522147

From MR I get this link as "Article": https://mathscinet.ams.org/mathscinet/leavingmsn?u=https://doi.org/10.1090/S0002-9904-1972-12971-9 - this includes DOI: doi.org/10.1090/S0002-9904-1972-12971-9

But much more often when I went to MR, I got simply a link to Project Euclid from "Article".

@LSpice Is it possible that sometimes DOI isn't listed simply because the publication doesn't have a DOI?

@MartinSleziak, quite possibly! I have to admit that I treat DOIs as slightly magic, or rather treat their existence or non- as somewhat mystical. So I have no idea why a publication might or might not have a DOI.

I am in the habit of just searching for the string "DOI" on an article page when I land there (or, as you say, getting the DOI from MSN if the latter provides it). I guess it could simply be as you say, that some publications don't have them, and that for some reason I am only encountering such publications on Project Euclid.

Thanks for the thoughts!

@LSpice I do not know much about ways of obtaining DOI. I would expect - as many things today - that it is a question of money. (Similarly as CC. But I would not expect DOI to be too expensive.)

Wikipedia article briefly mentions this: "Organizations that meet the contractual obligations of the DOI system and are willing to pay to become a member of the system can assign DOIs."

Thanks!