Will the classic IT department eventually disappear? It is always disappearing, parts of it at a time, but new parts keep appearing, or new uses of existing parts are defined. Take a look at the confluence of business process management, business rules and decision management. This is new "stuff" that needs expertise, even if its own goal is to increase the control of IT by non-IT business people.

The real divide is not IT versus business within a company; it is operations versus staff, the latter doing what is needed to keep operations humming, plus (very important) implementing changes to how operations executes. In the industrial age, machines were introduced, and in the information age, computers and systems were introduced.

In neither case did the rank and file of operations do this; staff people ran the projects that introduced change. Today, this includes a lot of IT people, irrespective if the change is implemented using in-house systems, packages, SaaS or outsourcing. These are the people who will introduce Innovations, the true driver of productivity improvement as stated in your related productivity article.

So, if you focus on the technology plumbing of IT, yes, parts of it will go away (not much call for machine code or assembler programmers these days) but use of IT (whatever its future form) in supporting innovation will continue.

David Wright
Business analyst, John Deere Credit Inc.

It seems Carr has taken a road of telling us what is on the horizon and telling us what the IT shop won't be. I see that horizon too, and I don't disagree.

But looking at an organization with thousands of employees, we can't have each individual employee act as an independent IT consumer without some guidelines or parameters within which to operate. We still expect product pricing to show some economies of scale. Those of us working in public arenas will still need to be accountable to the taxpayer for how tax dollars are spent on IT. There likely is a role for central IT as a gate-keeper, analyzer, synthesizer.

So my issue with the article is that I want to have that discussion about what we will become, what we want to be and what the IT shop of the future might look like.

Theresa Rowe
Chief Information Officer, Oakland University

In regard to Nicholas Carr's article, I think he might be taking the idea of Web-based utilities a little too far. If I were a CIO of any sized company, I would never be able to sleep knowing that my whole company could be brought to its knees because my access to the Internet is interrupted due to a faulty router or one of 100,000 other things that could go wrong with that connection. Considering things like disaster recovery, security, app customization, etc., it is more likely that the Internet infrastructure will buckle long before businesses will buy into the concept of moving their computing requirements and (even less likely) data out into "the cloud" in numbers big enough to affect IT jobs.

Lee Koenig
Project leader, Enterprise Rent-A-Car

While its premises are inviting, the claims stated in the listed book excerpt regarding how new applications will be a mix of universally available utilities oversimplify the actual internal workings of most applications. A

s an application development manager, who still programs(currently writing COBOL Web services and Unix scripts)my experience has been that the level of complexity required in the applications that our customers demand cannot be realized by such broadly available utilities.

As long as customers demand cutting-edge solutions that provide an advantage over their competition, the role of the IT department will never go away.

Chris Johnson
Application development manager
State Street Bank

Nicholas Carr really isn't predicting anything new; rather, I believe he is preaching the gospel according to Scott McNealy over at Sun. McNealy's been talking this up idea up for years. I believe that he may have coined the term "the network is the computer."

The Internet, being a giant public network, is then nothing more than a giant computing grid. So, having pondered McNealy's tag line for a number of years, I don't think that purchasing Carr's book will give me any more insight into this particular line of thought. Carr should read up on McNealy's thoughts on this topic...or maybe he has.

While Carr's thoughts on this issue are not without merit, securitizing a global network such as the World Wide Web will be a daunting task. Running reliable applications with confidentiality, integrity and availability is far more than the Web can offer today, if realistically ever. It's one thing to surf about, read e-mail and compose documents; it's quite another thing to run a large enterprise over the Web.

Running a large enterprise, whether you're doing it with pencil or paper or computers requires a certain degree of discipline and organization to be successful and competitive. Employees left to their own devices will all find a different way to complete a task, but if they're not using common tools to complete their tasks, the product of their efforts may not be intelligible to their peers within the same enterprise. Inefficiency and chaos takes over.

Certainly, IT departments will evolve and their makeup and mission within the enterprise will change with the tools that will be used to accomplish the various tasks that we are charged with. Until the computer, or the network for that matter, become as easy to operate as, say, a television, I don't think IT departments have much to get concerned about.

Mike Block
Assistant VP/IT officer
Equitable Bank

While Mr. Carr, a published author, is somewhat naive in his view of the future of computing, he does make a good point in that: the IT department is unlikely to survive in its current form. But he does seem to actually ignore the real trends in computing one can actually see.

He ignores the fact that someone needs to write the glue applications that enables all of the interconnection of "the cloud" and also whatever "next thing" that is still to come. This utility view of computing as being the future is a longstanding view of computing that has never materialized, nor will it. This is simply because there has not yet been invented a utility programming language for all of us to use to code what we want a computer to do.

From what I see from projecting current trends the future of IT does concur with a change in IT departments. It has at its core an elimination of mid-range computer platforms and mid-sized application development shops.

What we will see is an increase in the return to "glass house" computing environments. Mainframe computers running hundreds or thousands of virtual sessions servicing micro-computing access points that function as semi-intelligent terminals connected via secured wired or wireless Internet/intranet connections. This is because a mainframe-based environment provides the security, control of resources, power consumption reductions and per-user cost savings that businesses need now, or will very soon.

The micro-computing access point evolution can already be seen to be happening via the current evolution of cellular phones. However, it is my opinion that it will be the PDA form factor (or one similar to the larger MP3 player format) that will win the day by displacing notebooks, desktops, and cell phones as the all-in-one computing device. (This will occur due to the screen size limitation of the cellular phone format, wireless (Borg) headsets, voice recognition and cost. This transition will also not bode well for companies that make money from operating systems, since the underlying OS won't matter much in the micro-computing device environment.)

These new micro-computing systems will have limited actual capability for computing, being mostly designed for communication, storage and display (ergo the MP3 connection), with the real capacity residing outside the device in home or business data stores, all accessed via remote connections (wired or wireless). The devices would also function as one part of a two-part log-on system (hardware ID and a password or PIN) and could connect to other peripherals such as larger screens, keyboards, home theatres, etc., as required.

Therefore, in the future, I envision that general business IT shops will, for the most part, spend their time as application customizers modifying the outputs of purchased software (which could be either proprietary or open source) to feed to the "terminals" accessing those outputs as they wish it to be fed based on their access rights. There will also be a small group of component developers "building" new applications, but most of the real development will be occurring via a relatively small number of specialized software development firms or via group collaborations similar to that employed by Firefox and OpenOffice today.

Therefore IT departments will not go away, nor will they become redundant. IT departments will simply return to their roots and shrink in size, but not importance, as an integral component of either operations, finance or both. Jobs in IT will definitely become fewer, but comparatively better paid, and programming will return to the original "hacker" (a professional enthusiast) minority who will continue to be hired as consultants on an as-needed basis by business. While the rest of us will kludge things together using tools developed by others thinking we're programming.

The network is not the computer, just a part of one.

David H. Allingham, PMP
Environmental Protection and Climate Change Applications, Environment Canada

How can anyone disagree with this statement?

While smaller companies have strong economic incentives to embrace the full utility model quickly, larger companies will need to carefully balance their past investments in in-house computing with the benefits provided by utilities. They can be expected to pursue a hybrid approach for many years, supplying some hardware and software requirements themselves and purchasing others over the grid. One of the key challenges for IT departments lies in making the right decisions about what to hold onto and what to let go.

Most CIOs are doing this today. We have found very good utility companies for financial applications, e-mail and e-commerce.

Will our IT organizations change in the future? Only an idiot would say no.

But what we look like in the future depends on the value we provide today for our companies. If our IT organization is helping to uncover data relationships or improve business processes that will increase sales or reduce cost then that will continue in the future. But if we are not adding value then we should turn into dust in the wind.

Based on my experience with IT organizations and their capabilities to add value to the success of their company, my projection is a 50% survival rate. That's a lot better than zero!

David Bash
Director of IT, Nebraska Furniture Mart

I hate to admit it, but this time Mr. Carr is right.

As he said: "They can be expected to pursue a hybrid approach for many years, supplying some hardware and software requirements themselves and purchasing others over the grid. One of the key challenges for IT departments lies in making the right decisions about what to hold onto and what to let go."

My team is at this hybrid point now. Our five-year IT business plan has support staff focusing on becoming more consultant-like, technical staff becoming generalists and expert vendor managers, project managers becoming business process experts and our services coming more and more through the cloud.

Ultimately we will be the integrators of services and data and our value proposition will be our full understanding of the businesses we are in.

And, we will always hold on to our customers.

Dan Rainey
CIO, City of Ann Arbor, Mich.

IT is disappearing? Not at all, but it's a good topic with which to draw attention. The Internet is becoming our computer? What a shock! Go 15 years back and read Sun Microsystems' slogan: "The Network is the Computer."

Dr. Effy Oz
Professor of Management Science & Information Systems
The Pennsylvania State University, Great Valley
Malvern, Pa.

Hey, today I can buy something over the Internet, so tomorrow there won't be any more salesmen or marketing types.

But seriously, introducing new terms like the World Wide Computer is not really necessary. Let's just call them outsourcing vendors.

Large institutions like banks can outsource their core processing systems, since there are mature products and services for this large industry, with many standardized or regulated processes.

For other kinds of industries, large corporations will not outsource complex systems that give them a competitive advantage. They do outsource payroll and other systems that are non-competitive.

For outsourced services, your vendor must meet his SLA, and must be responsive to changes in the industry and customization request for each and every one of his customers. That's too much to ask for most industries and most outsourcing vendors.

It's a pipe dream to think that end users can turn a dial or flip a switch and construct and install major changes to an outsourced system, if that's what the author is saying. Plenty of IT people will be needed, to work for either the corporation or the outsourcing vendor.

Michael Tulig

Carr's observation are valid only in the most theoretical sense, and even his analogies don't stand up pragmatically.

1. Comparing computing to a wall socket

Carr assumes that the wall socket is sufficient for providing electricity. In many situations, it is the usual source, but is not deemed sufficiently reliable for critical use. That's why all hospitals have their own generators. Indeed, my local utility is sufficiently unreliable that I have my own permanent home generator, and I live 15 miles outside of Washington, D.C.!

He assumes that all power is of uniform cost and value, which it is not. There are many places where energy is not available "on the grid" but is vital. In such cases (rural areas, developing countries, etc), alternative means of generating local energy is more appropriate. Also, given technology evolution and the climatic impact of fossil fuels, it is likely that both cost and value of energy will change as alternative energy sources evolve.

He implies that the wall socket is the most efficient form of electricity. However, we already know that portable electricity is valuable, and becoming more valuable as an alternative to using fossil fuels in vehicles. Hybrid engines, electric cars, bio-diesel energy are all areas of active development.

He implies that electricity as delivered by the wall socket is how the electricity is used by the consumer. It's not at all uncommon for many electrical items to be powered, not directly from the wall socket, but rather, at a different voltage and current and therefy connected by a transformer.

If these pragmatic points are taken into account, we can say that for critical or potentially sensitive activities, local computing is necessary. Similarly, would could conjecture that not all computing capability (serial processing, massively parallel computing, distributed computing, database integrity, availability, and security) vary the cost and value of computing. Computing that is portable requires IT support, management, and integration (most large organizations have their own telephony departments!). And that, even if grid computing met most needs, there would still be a need for "information transformers" within some enterprises.

2. Creating a utility eliminates the need for local expertise

Accounting software hasn't eliminated the need for finance departments.

Statistical software hasn't eliminated the need for statisticians.

Certainly, accounting software has changed the nature of the activities of finance departments. They no longer hire bookkeepers. But the net effect is that finance is actually more complicated and creates a need for higher cognitive ability in its practitioners.

Similarly for statistical software, the average college graduate can plug numbers into a package and get mean and standard deviation, and can use statpacks for T-tests and Chi-squared tests, but definitive studies have statisticians involved who determine if the statistics are accurate (there IS an underlying mathematical model assumed to apply a statistical test), and who frequently are knowledgeable about potentially more powerful types of analysis.

Summarizing, Carr's approach tends to be thinking of information processing as relatively static and rapidly consolidating in its capability. In fact, grid computing, portable computing, digital media and platform variability are, if anything, demonstrating that we are moving away from a static and consolidative phase and towards an experimental phase, again.

Carr�s observations about several major trends that are shaping IT are thought provoking and right on, but I believe apply only to the immediate future. The accelerated interaction between IT and society is clearly one of the major drivers of unexpected and major change in individual, group, and organizational behavior.

Will all computing be in the cloud? Will IT become a commodity?

Possibly, but one wild card is the increasing use of mobile devices with better and better processing and less expensive storage.

Another joker in the deck is the growing problem of storing and accessing exponentially increasing amounts of data, which may lead to rethinking our entire concept of data storage versus intricate computing algorithms. Consider the historical trend in generating and distributing electricity. We've moved from small scale local generation to large utility conglomerates offering gigawatts of power via interconnected grids, to emerging energy conservation initiatives and individual solar and micro-scale generation feeding power back into the grid.

As long as information technology discoveries and engineering solutions keep evolving at an almost explosive rate, it seems most likely that computing in the cloud and IT as a commodity are only two stepping stones along a path of extremely diverse IT possibilities, including many unforeseen options that will continue to require IT specialists in technical areas that are not even recognized or defined today.

Jon McAdams
Executive Director, Client Solutions
iCohere, Inc.

Typical academic pundit. He is flying at 30,000 feet. He clearly has no actual IT implementation experience.

Several points:

In the better firms, IT is used as a �competitive weapon�. If everyone utilized the same utilities there would be no differentiation amongst the firms.

Secondly, Mr. Carr fails to understand many of today�s complex systems are beyond the capabilities of your typical utility. These generic providers can perform the basic functions. But the more complex requirements take time, customization and money. They are not capable or responding to complex and rapidly changing business conditions.

Lastly, any firm that has fiduciary responsibilities would be foolhardy to leave confidential data in an warehouse outside of their own premises

No one mentioned security. That little problem that seems to be getting bigger as we get more connected.

Companies are busier than ever putting up firewalls to keep thing out and in the process keeping a lot of things in also.

By I suppose this being a theoretical argument masquerading as a fact, in economist speak "we just leave that to solve it self."

Now if he could only write a book on that, it would be worth reading.

While I do not dispute the potential benefits of the shift that Mr. Carr is describing, he is ignoring (or unaware of) a huge impediment ... human nature.

People resist change!!! Even when there are obvious benefits. This includes the business users and existing IT staff. His prediction will
remain unrealized for a long time because of this resistance.

The average person will accept something brand new but they will resist changing their existing habits. As a result, change takes a long time within a given organization and it takes a really long time across an industry. In many cases, IT professionals resist change as much or more than their business users. The more people involved, the longer the change takes. Since Mr. Carr's vision requires both the business and IT to change, the chances that this type of radical change will occur in the next ten years is remote.

If you don't believe me, look at the continued dependance on mainframe computers and COBOL. People like Mr. Carr have been predicting the demise of COBOL for 30 years yet it is still quite pervasive. Don't get me wrong, I am not a COBOL advocate but this is a perfect example where industry predictions have not been realized.

Nicholas Spanos
Consulting Principal

Extrapolations have a way of glossing over impractical hurdles and so eventually fail to materialize!

Nicholas Carr's Book and your interview with him points out many flaws in his logic. Electricity is simple and stupid enough that once you decide to plug into the grid it is as simple as plugging in two prongs.

Corporate applications are not that simple, especially when it comes to many different aspects of corporate software that do the daily work. It is not security or comfort level of corporations having their data in some sight unseen servers. It is a lot more practical than that. Even the simplest applications require modifications or customizations that corporations require and a single grid may not provide these seamlessly.

Right after the Dot com boom, there were many startup companies that said "I can rent web server space from some company, I can rent hosted applications also the same way". Many of them raised tens of millions of dollars and struggling! Same kind of extrapolations, same mistakes!

There are parts of a company's IT that can go the grid way. If you need additional web servers to serve up your web interface if you are say UPS, can be done on the grid as demand requires. Even applications like Word, Excel can be served up over a grid with the actual documents residing locally.

It will be a long while before Carr's revolution will happen! It's not fear of technology but the simple practical issues that are usually glossed over that come back and make well made plans useless!

Nari Kannan
CEO, Ajira Technologies Inc.

Entertaining stuff but Nick should get out more... There remain many parts of the world with "last mile" challenges to say the least. And that's just the electrical grid!

On the 'cloud computing' front, I forecast:

* Fog and drizzle for Jane & Joe public but passable with care, wellies and a good umbrella.

* Intermittent sunshine and squalls for those with a business account overdraft to take them out of the stratus layers to about 10,000 feet.

* Clear skies and jet stream info tunnels for the gold card carrying super corps.

One wild card that might already be seeding this cloud concept with its own evolving mix of expectations is the younger generations taking their online world view into the workforce as both users and (good grief) as IT pros.

Mark McClure
CCIE #10814

Implementing information systems involves human activities which cannot be easily understood and standardized. If this was the case then we would not witness so many underutilized or failing IT systems. Carr's ideas are based on historical analogies with technologies whose usage does not require significant human involvement. This seems impossible when dealing with the development and usage of organizational computing probems. Unless we all agree on the features of the perfect IT expert to be cloned for the creation of the ideal environment of the cloud where humans would act always rationally without mistakes. We may instead just try to apply in practice more often a systems approach to IT and software engineering to understand better not just the technology part (which seems of predominant concern to Carr) but also the elusive human component of IT systems.