Friday, April 10, 2009

On the Cost of Macs

Elliott asked about my cost of ownership on my Mac.

I have to decline details for several reasons: I don't have the data from my 8yo Mac or the various software I bought; my software pricing will be different from others' due to large-volume site-licenses to which I had access; the purposes to which others will direct their machines will necessitate quite different software (and options) than I might (or did) use; and I have but an N of 1 for Macs that have gone entirely through their life cycle to death, and thus make a lousy source of meaningful conclusions (and providing a single data point for a later project of summing the numerous experiences of available bloggers will surely yield results skewed from whatever norm would represent one's own likely use).

Instead, I point out that Google can help one find "total cost of ownership" analyses depicting Macs (or Linux, or even OS/2 (which had its own zealots, and those who preached they weren't really zealots just sound thinkers)) against competition from Microsoft. (Heck, these guys offer a TCO study on how to run MS-Win applications most cheaply without actually buying MS-Windows machines.) One offers a TCO spreadsheet for calculating comparisons. The Unofficial Apple Weblog (TUAW) offers predictable fanboy praise of Apples as relatively cost-effective. Supporters of Microsoft's products' pricing tend to focus on up-front hardware cost. (Some question the comparisons on even the up-front costs, of course. Interesting comments in last link, incidentally. UPDATE: McCracken posted a comparison article that offers a number of feature comparisons before finally lining up spec-comparable prices.) Yet, up-front costs can be dwarfed by things like support costs and other overhead. This last link is interesting because it tries to address the arguable expensiveness of "free" software, discussing factors of interest to any cost-of-ownership thinking; it also contains this worthwhile reminder:
Both types of products have their pros and cons. It's best to pick the right product based on the project needs rather than viewing the product as a hammer, and all projects as nails.
Still, if one hunts I'm one can find specific enterprises have studied expected long-term costs of implementing different hardware or software solutions and have produced. Does one of these enterprises sound like it does what you do? Guessing the impact of software choices on your cost to operate isn't something that flows easily from buzz-words or ad campaigns, and products touted as creating savings can be as easily suspected of consuming resources (or at might result in a 'savings' that "excludes some migration costs and the price of software licenses" -- oops). Just figuring out your total cost of ownership turns out to be a white-paper subject. (Entertainingly, the site of the prior link offers on this page a few paid-for white-papers, including this gem: "This presentation, entitled Standardizing on Windows XP Instead of MAC OS X, provides a pre-packaged option for defending Windows XP against MAC OS X." In the old days -- well, after Windows 95 and NT hit the market at any rate -- you'd never see anyone dedicating the time to explain why to buy MSFT over Apple; it was self-evident that you would be buying MSFT, and nobody would dream of selling you an explanation for it. How the world has turned ....)

The real answer on raw price turns on what you require the machine to do. Many applications require software whose licensing costs (MSFT's server products?) or support overhead (anyone's relational database products for use in a custom database application?) dramatically overshadow the cost of hardware or operating system software. If this is your situation, the Mac/PC up-front cost debate isn't for you: it's immaterial. The real cost all happens post-purchase, and hardware pricing variance will be a mere rounding error. Some people have needs that are easily enough met on either platform that their purchase decision is made on quality-of-life issues involving their susceptibility to support by relatives, their need for tech support at all, the availability of previously-bought software, the capabilities of pre-installed software, or aesthetic aspects of the hardware or software. In some cases, the need for a waterproof computer or a machine that withstands dust storms restricts choices to one of relatively few hostile-environment vendors. In one comparison, the built-in ability of MacOS X to create PDFs from any application that supports printing turned out to make Apple's product cheaper than alternatives. What you need to do will dictate what you will need (at minimum) to buy.

What you want the machine to do will not only dictate its price coming into your hands, it will dictate the torture the machine must survive in use -- which in turn impacts lifespan, and thus impacts annual cost to solve your computer needs. My Powerbook that died last year had suffered drinks spilled directly into it on more than one occasion, which is why I had dubbed it the Aquanaut. Might this have lowered its lifespan? Who knows. Another reason my personal experience may not be extensible to readers :-)

Good luck, and don't get "taken" by shallow ads!

[NOTE: PC/Mac is a debate so old, and so fraught with religious tension, that the only possible defense against pure evangelism is perspective. Some worthwhile looks at the recent flare-up in the price comparison excitement include that at Fortune (including comments in both directions).]

UPDATE: Business Week offers a response to Microsoft's price-based ad campaign in an article titled "Mac vs. PC: What You Don't Get for $699"; the page with the meat of the author's argument is here.

2 comments:

Elliott said...

Thanks for the thoughtful post. I will check out the links you provided. It is probably time to think with a bit more care about the question of what I need the computer to do. Off the top of my head, having a computer that doesn't run noticeably slower within a year of purchase would be nice...

Jaded Consumer said...

Not running noticeably slower over time is something I've found Unix good for, whether Mac or not.

It's possible to drive machines to a crawl by overfilling the drives so there's no space for swap files, of course. Machines with modern memory management write information to the hard drive in case of later need to swap, for example, even if the machines don't actually need more room than exists to run loaded applications. However, just using the machine, collecting cache files, saving and deleting data in the filesystem, and changing various preferences settings don't seem to cause slowing over time.

If I knew more about how MSFT's operating system has evolved over the last decade (does it still have a fairly cryptic registry file to which lots of applications add information, to the potential peril and delay of system components that need to access the registry quickly?) I might be able to speculate more productively, but the Unix design theory tends to encourage applications to manage their own preferences -- and Apple's tools in particular encourage developers not to spray application components all over the filesystem but to have them all bundled in a single application bundle to enable drag-and-drop installation. The good part about drag-and-drop installation is that it is easy to install applications, and it makes most uninstallation extremely easy.

The propensity of MacOS X applications to create havoc for the operating system by replacing with custom versions components (like DLLs) relied on by the operating system is relatively low; each application can add its own versions of components inside its application bundle, and system components aren't stored where applications ordinarily write shared libraries. Applications have to be written with a special kind of "suck" to generate some of these problems, though of course they are not impossible.

Some of the comments on this blog addressing Cocoa (the native development environment for MacOS X, including the iPhone) illuminate some of the user experience differences between buyers of Cocoa applications and buyers of applications written for systems lacking features like runtime linking (which allows applications' component objects to "inherit" features from the current version of the operating system, regardless when the applications were originally compiled).

I think the investment case for Apple -- and I think it's a positive investment case -- is based in large part on Cocoa, and on Apple's ownership of its own hardware-independent operating system. Apple can botch services like the rest of them (read about MobileMe when launched; was an online service ever more snake-bit?), but Apple's ability to leverage good code into the future is a real asset.

Cocoa offers localization advantages, code reduction (through re-use) (and thus bug reduction, if bugs are proportional to code), resistance to obsolescence (OS upgrades enhance rather than obsolete applications), flexibility to target multiple architectures (and thus freedom from OEM supplier lock-in), development time reduction (Cocoa has been described as a RAD, allowing tiny shops to produce products able to compete with enormous competitors, which in turn helps attract innovative developers) ... the list goes on.

If you don't feel like picking up a new machine, consider picking up a few shares instead :-) I did.

As for the machine, many users will not describe technical advantages of Macs, but will instead talk about ease-of-use and other quality-of-life issues. These quality-of-life issues will change how happy you are when you work on the machine (if they apply to you; some may not); if you spend a lot of time working with your computer, you want to make sure you get the machine that makes this time least unpleasant.

Basically, it comes down to one question: How will you use the machine?

Answering that is probably the most important part of the battle, as it determines what you need to make your life with the computer a good one :-)