Old Software: a Problem or a Fact of Life?

April 16, 2013 Krzysztof Piskorski

When Windows XP debuted in October 2001, no one expected that twelve years later it will hold one third of the OS market. At the time, that was something unimaginable, just like expecting Windows 2.0 to still be around on Windows XP launch day (hint – it wasn’t).

How the things have changed! Today, Microsoft is literally giving money away to get people off Windows XP. Recently they have offered an 84$ discount for everyone who drops an aging platform and embraces bright-colored squariness of the latest installment. Why? Because, despite Microsoft’s best efforts, 38% PC users have decided they’re OK with old software and have shown no interest in any new products. Things look even bleaker in the enterprise, where Windows XP’s market share sits at a staggering 60%.

Faced with this fact, Microsoft is not only using a carrot, but also employing a stick. In this case, the stick is a death clock that started ticking away somewhere at Redmond offices last week. We’re now one year away from a moment when Microsoft finally drops the entire support for the system.

So, let’s get this straight: a company actively tries to combat its own product. How has it come to this?

Setting the recent slow-down in the PC evolution aside, it happened mostly because a product is only good if you can sell it to consumers each and every year.

I’m sure you’ve noticed this trend. Big software publishers are looking at various ways to increase revenue, and those always include shorter cycles, paid upgrades, limited lifetime. Many developers dream about annualized life cycles. Let’s take Electronic Arts as an example. Every major IP in their portfolio needs to have at least one installment a year, even though this practice, among other things, recently earned them a reward for the most consumer-unfriendly American company. Apple is also trying to sell regular MacOS updates, instead of a boxed version every couple of years. Windows is slowly heading in the same direction.

If you’ve ever wondered why Microsoft tries to subscribe home users to annually paid Office 365, by making the “boxed” license as expensive and limited as possible, here’s your answer.

Consumer electronics sector faced the same problem many years ago, but it developed a solution: engineered failure. No, I don’t want to talk about conspiracy theories involving products that are built to fail after a certain period of time. The real “engineered failure” comes in the form of a simple technical term called the reliability analysis. This means a product is designed to be sturdy enough to not fail within the warranty period. No one cares what happens after. If you can cut the cost while maintaining the same reliability window, you do that. Nowadays, the reliability analysis rarely reaches more than one or two years into the future. Do you remember a time when electronics often came with a 5 or even 10 year warranty? Today in US, Apple offers only 12 months as a standard, and its warranty practices were criticized by both EU and China, since they require longer warranties for consumer products.

The thing with software is that it has no moving parts, no grinding gears. It never brakes, it never wears out, and if it does, you simply reinstall a fresh copy. Despite the marketing efforts, it also remained almost entirely fashion-proof. If it wasn’t for various incompatibilities that accumulate over time, a piece of software could run indefinitely.

That’s a nightmare scenario for any business.

No wonder many companies try to actively dispose of their own legacy products. Some even try to incorporate engineered failures into their software. The rise of software-as-a-service (SaaS) made this a very enticing possibility. If your application depends heavily on a remote server, you can disable the server, and I bet Microsoft wishes they could do this with all the Windows XP copies out there. This is exactly what’s happening in the games industry right now. More and more recent releases, such as Sim City, are dependent on a remote cloud even when they run a single player campaign. This is not only a powerful DRM system. It also allows companies to “sunset” titles that have outstayed their welcome, and make place for new products.

This direction brings some new and unique challenges, especially in the field of culture. You can still play the original Space Invaders (a 1978 game) if you care to. But you can’t play the City of Heroes (a 2004 title) anymore – the servers were killed, the entire virtual world was erased. A long time ago multiplayer games allowed you to set up your own servers in order to play with friends. That’s not the case anymore.

Can you imagine a publishing house “sunsetting” Harry Potter in order to make a place for a new children’s classic? Of course not, yet that’s what’s happening in the software world!

It’s hard to shake a feeling that we’ve gotten worse at innovating, but way better at creating forced product cycles.

Personally, I’m in a peculiar situation. I work a lot with developers and I know how easier their lives would become if not for all the legacy systems, browsers and protocols they need to stay compatible with. I know they die a little bit inside when they see requirements for a Web app that include compatibility with IE6. I am also an Android fan, and I know how much it’s held back by enormous Android 2.3 user base that’s not going away any time soon.

On the other hand, PGS Software as an outsourcing company is happy enough to work with every solution or framework suggested by the clients, and to maintain their old platforms, even though we always try to suggest creative ways to bring old software up to date. In software development outsourcing, you don’t think about creating fake product cycles. You want to deliver a platform that will keep a customer happy for years.

I am also adding my bit to the old software problem. I have a small music rehearsal room down in a basement of my parent’s house. Over the years, I played there with many bands and the one sole piece of equipment that was always there was a Windows XP machine with some music editing software. It’s currently around 10 years old, and it has a pasty white CRT monitor. It’s there to record tracks, play samples and occasionally put on a YouTube video. It works just great in this regard.

This fact got me thinking whether software is more like personal electronic devices that I replace every two years or so, or maybe more like a car that usually stays for a fair bit longer than that? Or maybe good, stable software is more like a musical instrument? Like my Turkish-stamp “Avedis” cymbal, that’s at least 20 years old, and will be used for next 20 years?

Do you think my example – an offline, single purpose system – is a fringe case? Maybe, but I’ve seen machines like that in many small businesses. My local auto garage has a PC they use only to check e-mail and print invoices. And yes, it runs Windows XP. What will convince THEM to change it, short of hardware failure that might be years away?

If you take all of this into account, it’s hard to wonder why PC market is stalling at the moment. After reaching a technological plateau, there’s not that much functional difference between the Windows Vista machine from 2007 and a Widows 8 machine from 2013.

Consumer electronics industry tries to push clients to adopt new platforms every other year. Shortly after smartphones, we had netbooks, then tablets, and now insiders bet big on smart watches. But software is more impervious to fads, and that’s why developers try to give us other reasons to pay more often. The thing is, so far all of their he reasons sound more or less artificial. And, judging from the weak adoption of Windows 8, and by the huge outcry about “always online” games, the transition is not as smooth as IT businesses would have liked. Many developers also overestimate the “connectivity” of today’s world. While most people have access to a stable internet connection, the “always online” software and systems still cause trouble when on the road, on a plane, or on a vacation. For many buyers, having an option to run important software locally, without being dependent on the internet provider, is simply a must.

Next years will be very interesting in this regard. Either we, customers, grow to accept that in the  future software will be delivered to us only as a service, and that it will come with a hard-set expiration date, or the developers will discover that their dream business model is a bit detached from reality.

Either way, my small Windows XP machine loaded with legacy software will keep humming away in my rehearsal room, far beyond the next years’ expiration date. I doubt it even knows about this big doom clock somewhere at Redmond.

Last posts