The municipal transit agency in my town has been buying new buses since 2008, and it looks like they've just now gotten enough so that they don't have to run any of the old ones, at least not that I've seen in two or three weeks.
Although it can be easy (and tempting, even) to suggest that an organization should replace everything in one go, I believe the strategy taken by the local agency is an excellent strategy for an organization of any size, replacing any kind of asset that is somewhat expensive and has a life cycle on a rolling basis can help level out the budget and means that as the fleet ages and the oldest units die, there's unlikely to be a crisis where the majority of the fleet is unusable, either for a recall or because some part of the bus broke at an inopportune time.
What this means for the transit agency is that in a few years (when the buses they bought in 2008 are twelve years old) they can replace just those first few buses, as budgets allow, and continue on in this cycle.
The only problematic part for our local agency in particular is that they're (probably) also looking to expand their total fleet numbers for some additional services they want to run. The long-term plan is to change and increase service well through the next twenty years, which means they may need to come up with a way to purchase more buses some years. This creates a budget strain when the initial larger order of buses is needed, and each time that group of buses comes up for replacement, unless the agency is planning ahead enough that they can fill that need in advance. Essentially, they should be able to keep growing and refreshing if they buy some number of buses each year, even if that year isn't specifically a growth or replacement year already.
Buying buses you know you'll need in 2015 here in 2013 looks silly up front, but it gives the agency more vehicles to use to reduce wear and tear on each vehicle and gives the agency more time to maintain each bus. This is especially important if the agency is going to hire a mechanic, but only one, or is going to take the buses out of town for maintenance. This all of course depends on how the agency gets its budget, but it's presume for a moment it's a regular part of the city budget or that their operating revenue is enough to cover the cost of a bus or so.
The same thing applies to computers, and in kind, one of my projects is to determine the refresh rate of my computers and determine which machines need to be refreshed soonest, and, because I kind of bought a lot of machines in 2011 and 2012, what machines can wait to be refreshed while I refresh some of the older machines.
The other thing that's as important for me as it is for the transit agency is deciding what machines I need to have and what roles I need to fill. It would be nice if I could guess this in advance, but there are definitely differences between computers and transit buses that make this impractical. Not the least of these differences, of course, is that transit buses can be used for stuff even if they're not strictly mandatory, and that the requirements for transit buses change more slowly than for computers. I could not, however, have easily predicted my need or desire to own a desktop on which to play SimCity back in 2010 and bought something on sale then, expecting it to work as well as something bought today for the same cost.
Continuing along the thought train of what roles I've got computers in though: I currently have a server, a laptop (two actually, the primary and the previous), a Mac, a desktop, and a few tablets, plus gadgets and m phone. Presuming I replace the server, the desktop, the laptop, or the Mac each year; there's a four year cycle. What I've found, however, is that this may not be an accurate representation of the lifecycle of each of these devices, and the total cost of some of these machines is far larger (and incidentally, spread out over the course of months or years of building) than the other machines.
The other unpredictability factor in computer purchasing, at least as I know it, is that I may not remain interested in having a machine of each type every time it comes due for replacement. For example, based on a schedule I've developed, I'm unlikely to be able to buy a machine this year as I finally flesh out backups for TECT's large data volume, but I buy a desktop in 2014 and then a new Mac in 2015. Given that my current Mac is still fairly new, it is definitely a secondary or tertiary computer, and Mac OS X may be slowing down in system requirements, I might not need or want a Mac in 2015. I suppose in that case, I can either put money into upgrades for a machine I have at the time, or just keep the money on hand.
This brings me to my next quandary of premeditated computer buying patterns: should each type of system necessarily be refreshed at the same rate? Even though TECT is oldest, I have put it at the tail end of the cycle, purely because I know it's not going to be outmoded for what it's used for soon. TECT's job could almost certainly be done by topham, the computer I consider to be the least powerful among those I currently use. This means that in "the schedule" – TECT gets replaced with a new machine in 2017 or so. The biggest problem with this is that buying a machine as big as TECT (or bigger, who knows) is really expensive. TECT alone has handily passed over the most expensive computer I've ever bought, without backup equipment. We'll see if that particular equipment gets counted as part of TECT or as its own entity that can be moved from machine to machine.
The issue of replacing computers at a certain pace would be more clear if I were handling a bunch of nearly identical machines in an institutional environment, but it certainly is less obvious and clear when I'm talking about four or five machines in different roles. In the computing situation, especially at home, there are also things to consider like software licensing and peripherals. It's something I will definitely work on more closely as the need comes up, but the observation that the transit agency had kept buying new buses even though it looked like they had what they needed was interesting. In another year or so, I'll post more on whether or not they keep adding new buses before they expand the service.