The loss of HP, as it was from perhaps 1950 to 2000, wasn't just the loss of a brand or a manufacturer, it was the loss of an art form, a craft, a cherished part of engineering culture.
Their stuff was just so damn good, all of it.
A little detail that isn't often mentioned. In the 1980s or thereabouts, everything HP advertised was real. They never played the vaporware game, they never cheated just a bit on timing the ads. If you saw the ad in a magazine, it was finished, it was real, you could order it, it would arrive in a week or two--and it would work the way it was supposed to and meet all the specs. This, in a day when their competitors would run ads based on models or empty cases up to six months before the product was finished.
Using an incandescent light bulb as a feedback element in their audio oscillators was sheer elegance.
All their instruments were works of art. All of them had front panels that today's user interface designers ought to be studying. All the groupings made sense, almost every control was individually designed to perform its intended function. HP instruments looked good, felt good, were easy to use, and did exactly what they were supposed to do.
The first LaserJet was a revelation, and it worked perfectly, The first DeskJet was in many ways even more amazing--a 300 dpi printer for $600 when laser printers cost $3,000 and every other $600 machine was about 80 dpi if you were lucky.
HP's desk calculators were sweet, and the HP-35 was just a revelation when it came out. Everyone was proud of being able to do a square root, and here's this beautiful thing. Did everything a slide rule could do, everything, to ten-place accuracy when a slide rule would get you at most three. And, again unlike the competition--most particularly unlike TI--the math was impeccable, no glitches, no odd cases--they knew their numerical analysis and they got it right. RPN seemed weird, but at least it was consistent.The competition could never get this right--they would claim that you entered it "algebraically" but you would key in 30, then "sin" instead of sin(30).
The loss of the engineering days of HP was the loss of a whole discipline, a whole body of corporate memory on how to do things right. An irreparable loss of know-how. And it was engineering in the full sense of the word--these weren't self-indulgent overengineered toys, they were priced competitively and sold against competition in a real marketplace--and they were still so good.