I sort of feel like the osciloscope is a edge case in terms of test equipment. Its something thats really useful to everyone so I feel like its been hyper focused on for development.
But in terms of practical use, its difficult to compare a quality tektronix analog oscilloscope to something like a rigol 1052e. I find low end oscilloscopes downright annoying to use, plus you need to consider all the very advanced plugins that were made for certain scopes. With low end stuff you can get hit by firmware bugs.
I actually think that just now you can start to get really good quality digital scopes.. I think I also prefer the trigger system on older scopes.
There is also the repair factor, which can be a big thing, and the up time factor, if your scope breaks and its real expensive you can be in a world of hurt (say a 8000$ scope vs some 100$ old ebay deal).
But some other stuff, like power supplies, there is little reason for the modern ones. Something like a old lambda precision series PSU is a great value, and other then the fact that the display is limited, its very stable, very low noise and very cheap. I would actually consider it a questionable business decision to buy things like a 800$ wheras you can get like 3-5 high quality power supplies (between 1-3 channels each) on ebay for the same price, or even more, to have redundancy and ruggedness, high quality parts being used, established documentation.
I think that tools useful for digital circuit work have reduced in cost vastly, but still, I would never use a freaking bus pirate over something like a 1660 logic analyzer system, and there are even merits to using things like serial mainframes/mini computer things for some work so you have known good hardware that will work. Easy to access setup, quality cables, good UI, your never gonna be writing fucking useless code to test blink LED's if you have a setup like that.
This is a bit of a stretch, but take something like a PLUTOSDR that I bought. Extremely cheap, extremely good capabilities for the cost in terms of digital/communications work.. but I have had to struggle to use it, pretty much I need to run a separate with a version of linux I don't like to make the thing work, without too much work, otherwise I will be on linux forums all day.
All the minaturization and modern design trends are not all peachy IMO.
I for instance have some Sorensen power supplies for high current use, which are remarkably small and light for their power output but they are a complete bitch to fix, built super compact.. I actually would prefer to dedicate a space to a big old hunking linear supply to work with for the same power levels.. and they are pretty noisy (in both senses) too. I don't wanna deal with that shit when I am making a prototype really. I see it as something to use post development on a test bench after some high power subassembly is already tested in categorized, to approximate what can happen in a realistic power use scenario (where you might use something like a meanwell switcher of the right kind to power it). You end up dealing with more variables that can mask intrinsic problems.
I have a large 3A/8V power supply from HP. Made in 1960. All I ever needed to do with it was replace capacitor.. I see no reason why its junk or why I should replace it with anything else. I can take it off the shelf, put it anywhere I want and do something like a circuit burn in, remote module test, all sorts of things, with no effort to build a power supply, relocate primary lab units, etc. Boat anchor digital stuff often has weird input ranges that make it robust too. You can conveniently probe some weird signal level along with all the other ones without making adapters etc. If its some gigantic digital stream your gonna need to go with the poorly engineered cheap modern solution, but for nuts and bolt shit, keep it in a drawer. So much less frustrating.
But to be fair, I do have a fair bit of damaged/broke gear that is partially usable. But a economic and space analysis shows me that it siting on a cheap wire rack in the basement, with its cost, still is worth while compared to buying pricier equipment.. just the psychological factor is hard to get over sometimes. It mostly stays broke because there are fucking annoying ass design features which make it hard to fix (there is a HP function generator made of 4 PCB's all connected to each other in a cube made out of card edge connectors). Complete bitch to probe compared to some things. But I consider it a outlier compared to some other stuff. But I have also been getting side tracked alot and not planning it well. If I had to do it for a job, rather then enjoyment, it would all be fixed promptly with minor investment... I could strip and solder in the 40 or whatever bodge wires pretty quickly so it can be laid out on a bench and fixed, but I just dont wanna do that in my free time... if I got paid to do it though it would be done in a jiffy
with it being a hobby its just so easy to find other more entertaining things you can do once you reach a critical mass of crap...
***I do pitty the fool that does not have a digital oscilloscope