Despite the endless pressure to install the latest and greatest, many of the core technologies which are in use in the modern enterprise have been around for decades, if not centuries.

Our culture of newness means that people are embarrassed by "retro" technology. Nobody boasts that they've used the same fax machine for the past 20 years, or how Windows 3.1 still suits them just fine. (OK, we admit it, Slashdot types get a kick out of emulating ancient operating systems on their new PCs. But noone suggests that there's serious business value in doing so, and the very same people are the ones who are itching to rebuild their Linux kernel every time the wind changes direction.)

Innovation and freshness have become so entrenched as desirable business values that confessing you use older equipment is often viewed as tantamount to an admission of failure. Even with severe pressures on IT budgets, few people would come right out and suggest that we stop seeking newer, better ways of doing things.

Yet a quick glance at your own desk will confirm just how entrenched relatively ancient technologies are. The chances are that the PC before you is still running sections of operating system code which are more than 20 years old, displaying the results on a screen which is little different to those available a decade ago, and using a keyboard layout which was first thought up more than a century ago. You've probably still got an in-tray containing faxes, the concept of which dates back more than 150 years, and no electronic technology has yet proven itself a sensible replacement for the humble pen (the jury is still out on the tablet PC), which in one form or another has been with us for millennia.

Realising this can be quite a shock, but understanding why older technologies continue to thrive despite our relentless pursuit of the next big thing can help you understand where the real value in those technologies lies. It can also equip you to identify the crucial difference between systems that are out-of-date but functional, and systems which are actually costing you because of their age. So let's take a tour of some of the ancient systems littering your enterprise.

On the desktop

Operating systems
When Microsoft launched Windows 95 eight years ago, one of the big selling points was that it was the first Windows system to be natively 32-bit. Yet despite that shift, the software giant also went out of its way to ensure that much older DOS applications would continue to function. This wasn't ridiculously difficult, since so much of Windows was in fact just window dressing on top of a DOS shell which still firmly reflected its roots in 1981. It also reflected a market reality that Microsoft has long acknowledged: if you force people to upgrade their applications every time they upgrade their operating system, some of them will simply stop upgrading their operating system. This is rarely good for revenue.

Of course, Microsoft no longer guarantees that such ancient applications will work, and each Windows release from 1995 onwards has issued more and more dire warnings whenever you try while making the DOS prompt more and more obscure--but it is still there, and such applications will mostly work if you need them. (In fact, Microsoft has taken this process of nagging so far that many of the newest drivers from its partner companies get flagged by Windows XP for not being certified, but that's another story.)

Not everyone runs Windows, of course, but that doesn't necessarily make them any more advanced. Unix and its various derivatives (including Linux and Mac OS X) proudly acknowledge their heritage in decades-old operating systems, and indeed boast about this as one of the core factors in their reliability. The bottom line? No matter what OS you choose to use on the desktop, its historical heritage is likely to be at least as important as its future potential. That's not to underestimate the importance of some of the changes that have taken place--IP arriving in Windows, for instance--but you do need to keep a sense of perspective.

CRT displays
If you read through a PC magazine, or indeed a men's or business magazine, you could be forgiven for assuming that we were living in a world dominated by plasma screens. There's no denying that lightweight, low-powered, high-resolution flat screens (whether plasma or LCD) are indeed sexy. There's also no denying that relatively few people have them on their desktops, especially in business environments.

“No doubt, VOIP is the future but so is death—and you don’t see anyone hyping that.”

The reason? Unless you're planning to watch DVDs or play games--neither of which activities are heavily encouraged in most businesses--the monitor is one of the parts of your PC which can be recycled over two or three processor lifecycles. If your existing CRT monitor does the job, there's relatively little incentive to upgrade it. CRT screens are also actively included in machines aimed at the budget end of the market, or at specialised niches (such as Apple's education-targeted eMacs).

There are definite limits to this approach, though. As well as consuming more space, some older monitors may put out higher levels of radiation than are currently deemed acceptable, and many don't support advanced power-saving features. There's no reason to make your staff suffer by using a green-screen monitor, but the business case for everyone going flat screen simply hasn't emerged yet (and probably won't until several years after they become the standard option supplied with every PC, when basic apathy will again take over).

Keyboards
Save for the absence of control and function keys and a USB connector, there is scarily little difference between the keyboard you'll plug into a modern PC and that found on any typewriter from the past century. This is perhaps the ultimate example of training-based inertia.

The original QWERTY keyboard layout was designed not for efficiency, but for inefficiency; with a more logical layout, typists could attain speeds that were so high that old-fashioned hammer-style typewriters became hopelessly jammed. The need to jump all over the keyboard imposed by QWERTY eliminated this problem. (The French, of course, needed the differing AZERTY layout because the letter Q is more common in typical discourse.)

In the computer era, any keyboard should be able to keep pace with the fastest of typists (although some of the fold-out models designed for use with PDAs still have their moments). Nonetheless, QWERTY is now so firmly established as a de facto standard that nobody has ever been able to make serious inroads with more sensible keyboard layouts such as Dvorak.

Although companies with a concern for ergonomics may have updated their keyboards to minimise the risk of repetitive strain injury, there's been virtually no other change in this field.

While nirvana for many interface designers would be to eliminate keyboards in favour of other technologies such as pen or voice input, the flat reception to date for the tablet PC concept suggests that there's quite a way to travel on this front. And just as poorly-designed keyboards have survived most attempts so far to update them, they may well survive the challenge of voice and pen as well.

Processors
If you went to your PC supplier and tried to purchase a system based on the 386 processor, you'd be laughed right out of the supply chain. Modern processor technology now has such a high turnover rate that PC models rarely stay on the market for more than six months.

However, if you also do some digging, you may well find a staff member still using a 386--and if their demands don't extend much beyond checking their e-mail and producing the occasional document, that's unlikely to be a problem.

Notably, the fact that you can't buy a PC based on the 386 doesn't mean that Intel isn't still producing those processors for use in other applications. Older processors are frequently embedded in everything from toys to refrigerators.

In fact, the economics of building processor plants demand that they can stay in operation for several years. But don't expect your PC vendor to acknowledge this. (Indeed, Intel itself declined to discuss its embedded processor strategy for this story.)

Communication let me down

Fax machines
Although they didn't become widespread in Australian businesses until the 1980s, the basic technology for transmitting images over telephone lines was actually understood back in the 1840s. Once technology manufacturing techniques made fax machines affordable for most businesses, they quickly spread into companies of all sizes. In the pre-Internet era, faxes seemed like the ultimate way to communicate documents between businesses, even if you had to photocopy them for archival purposes and curse at the relative illegibility of many of them.

In a post-Internet world, of course, things are rather different. Being sent a fax these days can seem like an inconvenience compared to the flexibility of e-mail. Yet fax has stubbornly hung on in there, and continues to be used even for applications which might seem more suited to an Internet-based approach.

The explanation, again, is procedural. Businesses that rely heavily on paper documents--such as legal and financial companies--very quickly adopted faxes as part of their overall workflow. Shifting to a purely digital approach is much more complicated, and if there isn't a clear benefit to doing so, the shift simply doesn't happen. The lower archiving costs for digital documents versus paper ones are likely to tell against fax systems in the very long run, but that change could be decades in the making.

Telephony systems
On the surface, it might seem that telephony systems have changed most of all in the past decade or so. Nobody uses analogue phones any more; many people have shifted away from conventional telephony to IP-based systems, which (at least in theory) allow voice applications to be merged with all kinds of other software. And then there's mobile phones everywhere you turn.

However, while the underlying plumbing may have changed significantly, the uses to which telephones are put haven't undergone the same kinds of fundamental shift. For instance, integrators universally report that the main reason for using IP telephony systems to date has been to cut down the costs of existing calls. That doesn't mean that the people implementing those systems aren't aware that they can also be used for neat presence-based applications and the like, but no-one appears to be making the shift for that reason. As a result, the only indication many users will have when they switch to an IP system is a new handset, and once they start making calls, they'll probably forget about that too.

Backend to the future

Tape backup
Ever since the first disk drives were rolled out into businesses, doomsayers have been predicting that tape would one day give way to disk as a major backup medium. The continued collapse in storage costs on disk (which is now approaching just cents per gigabyte in large volumes) and the emergence of new technologies such as iSCSI means that this prediction continues to reappear every few years. Nonetheless, there are very few enterprises which don't rely on tape as a primary element in their backup strategy.

Tape's easy portability and relatively low cost for larger volumes have ensured it an ongoing place in the enterprise arsenal. While capacities have increased, many of the fundamental mechanisms in use have stayed more or less the same over the past two decades. What has shifted is the way in which recovery is performed. While tape was once the medium of only resort, short-term backups to disk are often now used in the event of emergencies, with tape reserved for more fundamental restores or for use in unexpected natural disasters.

The shift away from tape is happening gradually; for instance, the most recent release of Oracle's flagship database is the first that assumes by default that backups will be performed to disk rather than tape. Despite such advances, though, nobody expects tape to disappear for a long time yet.

Mainframes
Mainframes are to the IT world what tortoises are to the natural world: they've been around for a long time, they never seem to die, they're kind of ugly, and they continue to do what they were designed to do originally. Oh, and they have a weird kind of shell interface.

Even though processing power has skipped ahead in leaps and bounds, mainframes have survived several generations of IT fashion simply by performing basic processing tasks with a high level of reliability. The current trend towards server consolidation suggests that users are again warming to the notion of centralised, virtualised, highly powered applications--the very concept around which most mainframes were designed.

While no company would base its whole business strategy on new mainframe sales, those which are in active service are unlikely to disappear in the near future. It's safe to assume, for instance, that any mainframe application that it was deemed worthwhile going through the hassle and expense of updating to eliminate any millennium-related bugs is going to be with us for some time to come.

Web services
The concept of Web services--effectively miniature applications which use XML as a standard data format and allow enterprises to combine together individual tasks to match their business needs more precisely--was first promoted as a way of bringing together highly customised components from a wide variety of sources. Yet their greatest value to date has been in providing a means of integrating disparate enterprise services. This in turn can eliminate one of the biggest headaches in any enterprise: the challenge of integration. (Analysts often estimate that integration accounts for 40 percent or more of the typical IT budget.)

While Web services have enormous potential in providing means of linking together disparate applications--some of which may be senior citizens in IT terms--the process isn't an automatic one. Most applications will require at least some modification in order to provide output which can be used by other Web services. Newer releases of most core enterprise applications often have Web services adaptors built in, but if you are working with older systems you may need specialised help. On the upside, once those adaptors are in place, you may be able to minimise maintenance on those core applications.

Not ready for retirement

Making solid long-term use of existing technologies makes good business sense, since it can reduce your overall IT spend, or leave you with extra dollars for innovations that can help differentiate you from competitors.

Think value, not vintage. The notion that technology needs to deliver real business value is a familiar one, but is very relevant in this context. If a system is offering real value, don't ditch it simply because of its age.

Make sure there's a support chain. No matter how effective technology is, you need to be able to fix it when it breaks down. If there's no-one left who can offer service and support, it may be time to move on.

Get good documentation. On a related note, make sure that the procedures associated with any technology, new or old, are thoroughly documented, so that you're not reliant on the knowledge of a single individual. This is especially important with custom developed system.

Integrate with other processes. As long as existing technologies can work effectively with new ones, there's no reason to ditch them. If a single system is hampering efforts to streamline, then you may need to think again.

For whom the fax polls If you wanted to conduct a poll across a large and diverse group of people, chances are you'd think of implementing an Internet-based system. Yet one of Australia's largest public voting systems runs not on IP, but by using faxes. ASX Perpetual Registrars, a joint venture between the Australian Stock Exchange and Perpetual Trustees, handles voting by shareholders on a huge range of matters relating to listed companies. Those voting documents are sent to shareholders on paper; each individual can choose between returning them by fax or post. Such a system ensures that noone requires anything more than access to a mailbox to vote on company matters, but obviously imposes a huge processing workload. Entering in all the data manually not only involved significant re-keying of data, but also introduced the possibility of error and could delay the vote in crucial matters such as board elections. The company solved the problem not by abandoning paper voting systems, but by providing more intelligent processing at the receiving end. The combination of an Esker fax gateway and a Readsoft text processing engine allows faxes to be automatically processed as they are received, and the predictable format of voting forms means the accuracy rate is extremely high. Votes received by post are scanned and then passed over to the processing system--a system which still requires manual intervention, but significantly reduces the potential for error and provides speedier handling.

Subscribe now to Australian Technology & Business magazine.

Copyright © 2009 CBS Interactive, a CBS Company. All Rights Reserved.
ZDNET is a registered service mark of CBS Interactive. ZDNET Logo is a service mark of CBS Interactive.