Ask HN: Perceived vs actual computer slowdown, over time

5 points by danpat ↗ HN
The question of "perceived slowdown" came up at work when debating the costs of upgrading some desktops vs wiping/re-installing.

I'm sure everyone is familiar with the delight of a new, snappy computer. However, 3 years down the road, the computer inevitably "feels" slow.

I'm not looking for advice on removing malware, cleaning up filesystems, optimizing disk fragmentation or any of the other commonly talked about remedies. What I want to know is whether the phenomenon of the slow-computer-that-used-to-be-fast is a real thing or not. Is it our expectations that are changing over time, or is something else happening?

My Google-Fu reveals nothing relevant here, so I turn to HN. Does anyone know if studies have been performed that compare actual PC performance vs perceived performance and how this changes over time?

10 comments

[ 4.6 ms ] story [ 43.9 ms ] thread
I don't know of studies, but my anecdotal experience suggests that computers do actually become less responsive with use, and that reinstalling the OS restores them to their original performance.

What I am almost certain happens is that computers accumulate cruft (registry entries, competing installed software, forgotten preferences) that slows down what seems to be normal operations for the computer. This is far worse with people who don't know anything about computers. I just saved my in-laws from buying a new computer because they thought the old one was wearing out. Turns out they had installed two anti-virus suites, which is enough to bring just about any machine to its knees.

If you had a machine you locked down and never installed anything on it, there's no reason why it would slow down gradually over time. Computers just don't work that way.

Agreed. What I'm looking for is a study that quantifies some of what's going on here. How much actual performance is lost over time, how quickly does it happen, and how are both of those factors perceived?

The case of your in-laws sounds pretty typical, we've all heard it. Purchasing a new computer seems like a rather dramatic solution, but in their minds, fairly justified.

I guess another question in my mind is the opportunity costs involved. Is the time spent restoring a computer to it's original state worth it when compared to simply purchasing a new machine? Of course, this varies with the ability to quickly re-install, so I'm considering this question mostly on the small end of the scale (home users, small businesses, where economies of scale haven't kicked in).

The only OS that I've encountered suffering from this problem is Windows. Both OSX and Linux manage to retain the same amount of perceived responsiveness to me. It's all anecdotal.
The slowdown is real if you keep upgrading software. I have a 17" MacBook Pro 2,1 from 2006.

It shipped with Mac OS X 10.4 and it's running Lion now. Every upgrade has taken a little pep out of its step.

Software keeps advancing to reap the benefits of faster hardware.

Back in the 1980's I ran a ray tracing program called "Silver" and thought nothing of waiting 20 minutes for a screen resolution rendering of a couple of mirrored balls over a checkerboard. Now I expect complex objects to render in high def in real time (o.k. maybe not with ray-tracing but with shadows) and if I wait 20 minutes for a rendering it will be at many megapixels and use radiosity.

But your real issue probably isn't one of perceived performance, it's one of perceived technical sophistication. My smart phone is many times slower than my desktop, but it's new and cool and the desktop is over three years old and kind of boring - and over the wifi network they are both about equally fast.

So the question is, is that phenomenon of "perceived technical sophistication" and expectation something that has been studied.

I'd be really interested to know how people generally behave here. Questions like, "use X for a while, then get to try out X^2, then go back to X, how long does it take people to forget the speed improvement of X^2"?

In a workplace context, I don't think that's even the right path. For some people, the purchase of a new computer represents the idea that the company cares. In other words, it is a work environment issue - it is more pleasing to be around shiny new iMacs than three year old Vostro's (so to speak) - even if the iMac is actually slower.

Of course - in a typical work context (i.e. outside tech and startups) there are lots of people who would prefer to use the old computer and get a bonus approximating the purchase price of a new computer - again the new computer in the workplace is usually a social issue not a technical one.

I think once SSDs become standardized we'll start seeing a lot less "preceived slowness." Disk access always seems to be the most noticeable issue after a year or two of use for me.
A 2008 computer running 2011 software will be more heavily loaded than a 2008 computer running 2008 software, which will be more heavily loaded than a 2011 computer running 2008 software - hardware gets faster, software gets less efficient/does more.

As to "OS slowdown": I have, in the past, read posts that say that reinstalling Windows made their computer fast again. Whether that's due to less registry cruft or removal of unnecessary programs, I couldn't say, but it does suggest that the effect is not completely based on hardware getting better or new versions of software using more resources.