Agreed, also with OP. I'm not sure why Alex didn't mention that the original publication date (or date range) is a standard item in the `content.opf` file that is part of every ebook.
This data is important to me, too, when choosing ebook—e.g., reading Dickens or some other prolific author in order of publication. The edge case for non-inclusion is a bit of an outlier–itself readily solved by using range, as noted (and producers already can do a fair bit of research when producing an SE volume!). (Conversely, the date of publication of the SE edition is of no interest to me at all...)
For personal interest, I maintain a "cumulative index" of the SE corpus which includes the original publication date, scraped from the `content.opf` file. Anyone interested is welcome to have a look: https://www.sudalyph.org/seci/
I'd push back on your example and say it doesn't matter if you say 1597, 1599, or 1609—any of those provides quite a lot of information to the reader and vastly more information that not providing any date at all. I find it a bit hard to scan because some of the books seem to be from the Roman era, others 1800's, and having an approximate publication year would be very helpful.
But again, I could be wrong, just suspect OP is right and putting years frequently would be a decent improvement.