It seems like one of the more difficult questions when it comes to computers is when, exactly, you should make a purchase. This is all related to factors other than budget which make people wait to buy a new system. I have always had a challenge with the decision on when to wait and when to purchase a new thing. This type of decision applies to everything, but (with a few fairly old systems) I've been facing this particular decision myself. Of course, this is only one factor in the overall decision, but the argument that "something better is just on the horizon" does come up.
If you go too early, you might miss out on new functionality, a big performance boost, or a specific configuration. If you wait too long, you'll be missing out on using whatever the new piece of technology is, and there will be the costs associated with that in terms of, however much profit or "productivity" you could have achieved with the new thing. (For example, how much World of Warcraft couldn't you play because you were using a seven-year-old laptop with unsupported graphics?)
I'm currently in this particular waiting game with most of my computers. It's challenging because I have been having a reasonably good time scavenging parts out of old systems, and that in and of itself can be fun, and rewarding to a certain extent. I've also been struggling with the fact that even though the system may not be ideal for some tasks (World of Warcraft, for example) most of my systems (including the scavenged ones) aren't "over the hill" yet.
The other aspect I discussed in the Computer Lifespan post is that certain parts and systems have a physical lifespan as well as a lifespan at performing some tasks. Even if a computer can still run Word, and even do so quickly, the hinges on some of my systems are old and bezels and other plastics are wearing out, some systems are growing out of their RAM ceiling or need yet another replacement disk or some other part.
Some of these things aren't strictly issues for scavenged systems, but there are both decisions to be made about how much money to spend on the scavenged computers, and at what point you've spent too much on a scavenged system. This issue is two-fold, especially given that today's Very Inexpensive Windows Computers are getting better and coming down in price.
The other aspect is at what point are you waiting too long for a new system? One of my recent hobbies has been comparing superslab, my ThinkPad T400 and my current main computer, to the Microsoft Surface Pro 3, a system that's arguably an improvement in every way, and would be functional both as my main computer at my desk and on the go, replacing my Microsoft Surface RT. This commentary applies to any potential new desktop computer just as well, however.
One of the pieces of advice I've received recently is to wait for Intel Core M to drop, which should enable the Surface Pro in the same or thinner chassis, but without a fan, or even thinner and lighter notebook computers, with reasonably good everyday performance. The other advice I then received was to wait for the version of the Surface or Surface Pro that ships along with the next major version of Windows. Of course, at that time, we didn't have any information about when, exactly, that version of Windows would ship. We know today that it may not be until late 2015 that it does.
Windows versions affect desktops slightly less than they do portable computers, because a portable computer is likely to perform best with the version of Windows it shipped with. (In my particular case, the ThinkPad T400 has become less and less well-supported and less and less mobile overall with each successive release of Windows.) However, because it gives a specific target for physical features, performance, and functionality, I think that buying both desktops and portables based on hardware timelines has merits. The problem is that Intel publishes information about its timeline well before anything ever ships, so it's possible to know, for example, some of the properties of the Skylake micro-architecture, even though Broadwell hasn't shipped yet.
Another thing I worry about is whether or not waiting too long may result in getting a product that isn't likeable, or the discontinuation of something you were hoping would get a successor. As I mentioned in a recent post about it, I like Windows RT a lot and have concerns that it may be discontinued as a product or as a platform I would like to (even if it's not strictly necessary) buy another RT device before the platform dies, and the decision of whether to buy a Microsoft Surface 2 or Nokia Lumia 2520 now or wait until a next generation device maybe comes out is one of the more challenging ones. (In particular, from my vantage point it looks like Windows RT is primed for a much more powerful processor and more memory, such hardware is available, but it simply hasn't come out as a product yet.)
Another example is that many people waited to buy Lenovo ThinkPads based on the Intel Haswell processors, and later expressed disappointment that they didn't buy a model with a Sandy Bridge or Ivy Bridge processor, due not to an existence failure of the product they wanted, but changes in the physical properties of the product. In particular, Lenovo changed the layout and form factor of some of the input devices in these products.
Often, people will say to buy what you need or want when or need you want it. For mostly everybody out there who isn't as specific about this kind of stuff as I am, that's really great advice. Most people can tell when they 'should' get a new computer, and to tie it back to something I mentioned in the Computer Lifespan post. I think most people aren't nearly as heavily invested in their computers, either because they're using computers that an employer paid for, or because their computer buying process equates to going to a local retailer and kicking the tires on machines until they find one that looks okay, seems fast enough, and doesn't cost too much for their individual budget. It's not a bad way to buy computers, but it's very uninvolved.
Ultimately, even if I had the budget, I'd be concerned about when to make the move. Contrary to the popular opinion, I remain interested in maximizing the use of any given machine, which is a big part of why I do my writing on a Microsoft Surface RT long after many people have stopped using theirs, and why I'm still playing World of Warcraft on my ThinkPad T400, even though it is no longer supported. I have the scavenged machines, but given that those have (or could have) their own issues, I'm not counting them as part of my purchasing cycle.