Well, if you are talking about the symbols printed ON the keys themselves, there is a sound economic reason for doing this.
If you have to do double-shot precision molding, that isn't really cheap, but the more of it you do using the exact same tooling (and thus legends), the cheaper it is per unit. So you anticipate the total eventual demand for a product, add a percentage for replacement and repairs, go to your plastics guys, and have them mold a heck of a lot of key assemblies. You warehouse them until needed in production.
Or, you decide to print the keytops instead, since printing is cheaper per key. But to make key legends durable, the printing must be thick and hard-- the usual ink/paint for this is an epoxy-based paint that is as hard and tough as the plastic beneath. Smaller runs of this is okay, but the process is very variable; to assure color consistency and clarity and durability throughout the product's run, you'll make a lot of them all at one time, then warehouse them until your assembly line calls for them.
But your wonderful product may not sell quite as good as you estimate. Ooops.
Eventually, the Marketing department determines that what is needed is a new model of calculator, with different symbols and functions on the keys. You COULD just redesign case and keys and such from scratch, going to all the expense you did for the prior calculator that didn't sell. OR...
You can ask your design team to reuse some of the elements of the prior design. This saves a lot in the design, tooling, inventory, production startup costs. It is also apparent you'll NEVER get rid of your prior parts inventory unless you take some of it to put in your new calculator. Why write off and destroy what you can re-use?
Keys are easily "reused" in a design, except for one minor detail: the symbols are all wrong for your new device. But there is spray repainting and then new printing, which allows you to paint and re-legend the old keys to your new specs. No, it doesn't look as great as a "fresh" keytop, it has less color consistency, and isn't as deep or hard on the surface as an original pad-print on a blank key-- but it solves so many of your problems, and users just don't seem to care as long as the product meets their needs and is cheap enough to buy.
That, Bill, is what you are seeing. Manufacturers face a dilemma: not every product they focus-group and test market and then release will actually be met with marketplace success. Then you cut your losses, re-design to follow market trends, and use the experience (and sometimes even the parts) of the failed product as leverage to be faster rolling out the Right product, have it cheaper, and have the features the public desires.
Repainting and subsequent pad-printing is a standard activity of manufacturers of keyboards and keypads. What they do it OVER, depends entirely on what they already have as molded and/or painted key stock.
It is maybe a problem that HP was so good at their keys for so long-- on much of their product, they designed their keypads as elegantly as the rest of the machine, sparing little expense to make them durable and cosmetically attractive. So you REALLY notice when HP cuts a corner here.
TI (and many other manufacturers that HP has felt it had to compete with) rarely went to such troubles; their keystock was typically blank, their legends ALWAYS epoxy pad-printed in "more-realistic" runs since they aimed to make many different models along basic design "families", and they knew their product was not dependent on cosmetic consistency as much as its functionality and price-point. Maybe also for a while, HP was too optimistic estimating its market, in the face of gaining competitors.
So as HP adjusted more and more into their "Be Competitive" philosophy, you have noticed that some keys on your "new" calculators are "leftovers" from excess runs of old stock.
If you feel slightly *cheated* that you got some leftovers in a calculator you paid so much for, you might check the list prices of the models that preceded it. To keep up with the dramatic change in the economics and fashions of the calc world, HP (in a tumultuous period) exercised some miserly options.
Your calculators are good ones, and will ALWAYS be highly valued for their utility. But few would argue their cosmetics, finish and durability were as high a priority for HP on those calculators as those of prior "generations".