Adding a lock-in to prevent the user from using a different vendor for battery packs is not acceptable, but I think authentication of the battery might be good anyway...
A device manufactufer is good at making devices, not batteries. There is no reason why a competitor may not be able to come up with a better battery technology, especially if that competitor is specialized in energy storage.
Not letting the end-user "upgrade" to a better battery pack would be wrong: it's a simple antitrust issue.
As for the users wanting to blame the device manufacturer when using a cheap knock-off battery, warning the end-user of the loss of guarantee should be good enough, be it by putting a sticker on the "real" batteries or by displaying a warning when you boot your device with a knock-off that doesn't pass authentication.
This is why I think a full lock-in is overkill, but authentication can help the end-user put the blame on the correct part-vendor when their knock-off battery fails.
Now, it's up to the device manufacturers to decide if they want to be evil or not.
Ahhh, NiCds--my ancient TI SR-50 scientific calculator (ca. 1973) used three AA-size NiCds. When they evenbtually failed, I was able to buy replacement batteries for a few dollars and place them into the cell-holder of the calculator.
Of course, between the high power drain of the LED display and the modest capacity of the cells, run time was not great by today's standards--but hey, it was wonderful at the time!
I may not always like it, but I agree with the idea of a proprietary battery. You are discounting physical properties. The physical and performance properties of lithium chemistries vary widely and even widely within a particular chemisty. Because of the way things work, people are quick to blame a manufacturer for battery problems or mishaps even if the battery was a knock-off rather than a manufacturer-approved battery pack. So it sometimes makes sense to exclude packs that may not be up to snuff.
Also, lithium-ion cells are not cheap. That is why battery packs cost so much.
This idea of proprietary batteries is just wrong. Just like with printer cartridges, they are a marketing gimmick. True, there are technical issues (correct ink or battery physical properties) but they are simpler than the proponents of proprietary pairing make them out to be. After all, semiconductor manufacturers can design battery control ICs that work with wide variety of cells.
The prices of proprietary batteries are just excessive---come on, you can get a 20nm, billion transistor CPU for under $10, and the battery pack costs $100? I do understand that there are unscrupulous vendors that cut corners and make substandard or even unsafe packs, but the reputable vendors should concentrate on making safe, inexpensive power sources---rather than on gouging the customer via proprietary lock-in. After all, as the title of the article says, Battery Growth Drives IC Opportunities.
Besides, proprietary lock-in can always be technically bypassed, and the courts decided that it is a legal reverse engineering.