CTO Blog

A PDA with mobile email does the trick in most of the cases for me :-)

There should however be some work done in making these web apps (saas whatever) suitable for those devices.

Lee

Yes, the desktop is critical. What happens when the net isn't there, unavailable, or expensive. Applications & Data have to be available with zero to a heavy footprint. I don't know what the right answer is, but I do know that when the net isn't available, I still want to do "stuff". Regardless of what I have to work with -- cell phone, berry, pda, terminal at a inet cafe, notebook, or desktop -- I need to get things done. Argh!

Yes it does and refer http://www.gandalf-lab.com/blog/2007/04/question-is-does-desktop-matter-anymore.html for why I think so.

Peter,

The recent Knowledge@Wharton interview with Ray Ozzie (http://knowledge.wharton.upenn.edu/article.cfm?articleid=1698) gives a good illustration on what does the biggest OS vendor thinks about the balance of about local v. web applications.

A possible middle ground you are talking about could be the use of corporate virtual machine builds on employees home laptops. This will not move us to a 'thin' desktop, but it has few other benefits: gives capability to work offline, reduces the number of devices user has to worry about and decreases in capital expenditure related to hardware inventory, hardware support and maintenance cost.

In regards to the cost reduction, I would be quite interested in economics of most current hardware lease back and support contracts : In theory the large desktop hardware contract should be cheaper thanks to economies of scale etc. Yet my suspicion is that there are probably quite large overheads which are passed down to customer, so the end result actually may not be that great.

I’ve been playing with the virtual machine solution, but it seems like a short term fix while we sort out the new role of the desktop since it ignores that fact the boundaries between organizations are eroding.

At the moment we worry about managing our users' desktops, however the real problem is that the distinction between our own users and our partners, and even our customers, is starting to blur. If I hire you on contact for a few months then am I going to give you a managed laptop? Why would I when you probably already have one that works for you, and you don’t want to be burdened with another device? Can I put a virtual machine on it? Does the VM I use have a version that will run on your Ubintu instance? Do I support the VM when it doesn’t work on your device? How do I collaborate with partners and clients? Do I provide them with my fat clients (via a device or VM) too, so that they can get at the stuff they need?

We're moving to a world where companies live inside larger ecosystems, with people moving around the ecosystem as need be. A good example is product design, where it’s now common to pull in partners across the entire value chain in the design of a new product. I can’t mandate a single common platform across my partner and customer ecosystem, but I can deliver some functionality over the web into a device you probably already have. In this world the web is the lowest common denominator, and if I’m delivering an application to clients over the web why wouldn’t I use the same channel internally? Couple this with the fact that the advantages of the "rich desktop experience" and "ability to work offline" are rapidly disappearing, and the argument for business owning the desktop starts to have problems as we don’t need the 100% solution it provides. The 80% (and growing) solution I can provide over the web is enough.

However, the real killer for me is the shift from channels (email, word processing, IM, backend applications, ...) and places (my desktop, our partners office, the client site) to shared spaces (Basecamp, Aconex, …). The new work paradigm that's emerging is to create shared spaces to marshal a team around, and then pulling team members, technologies and references from across the entire ecosystem into the space as needed. This promises to turn our application centric approach to IT on its head as we begin to treat applications as information appliances, responsible for maintaining transactional data in the background and delivering it into the collaboration spaces that we create. Spaces provide a single shared point of reference with in the business (to answer a FOI request, to manage a case, as a shared planning environment, …), with the potential to create significant cost savings by ensuring that all stakeholders (both within and outside a company) are informed, involved, and have the information they need to get the job done, driving up both quality and productivity with an immediate impact on the bottom lone. The desktop becomes a channel to access this space, and if the desktop is just a channel, like the mobile phone, then ultimately its relationship with the business will change.

That said, we’re not talking about the end of the desktop. We will still need one, but it’s becoming a personal device. I want my laptop, set up the way I like it, with my productivity tools (you can pry MindManager from my cold dead hands), and I’m happy to bring it with me to the job. The business will only care that I have one, but other than that it doesn’t matter.

Peter,

I agree that what you describe is the direction things are roughly heading to. At the same time, I think there are some important constraints to the feasibility of the vision. I believe that this will affect timing and scope of implementation of the vision.

Reinkefj post points to connectivity as one of the factors critical for the vision to become true. There is a clear trend that large parts of workers, especially in knowledge-intense and creative industries, are becoming much more mobile. Yet mobile knowledge worker scenario poses some of the biggest technical challenges to the realisation of the vision, which in turn provides a good test to assess the extent to which the vision is feasible.

Mobile knowledge workers want to get on with their job wherever they are. This means a requirement for ubiquitous access to their data and applications - not only in the office, but also on the go. I can broadly see two technical means how this can be provided to deliver the vision you describe: one network based and one application based.

One alternative would be to use thin client applicatons. This however require bandwidth to become so cheap and ubiquitous that people will not have to worry about it. Let's make a small though experiment at this point: if your employer offered you access to all services and data you currently have on your laptop via mobile GPRS-enabled dial up and Citrix, explicitly excluding the possibility to store data locally, would you take up him on his offer? I certainly would not.

Despite the availability of bandwidth steadily improving, there are still plenty of blind spots without connectivity - be their rural areas and public spaces - which preclude true on the go connectivity. Apart from that the way service providers charge for wireless service puts some practical limits the time you are willing to spend connected.
I am guessing it will take few years before both of these factors will change and connectivity will becomee sufficiently cheap and ubiquitous to allow the provision of most applications in a software as a service manner.

Other alternative would be not to depend on bandwidth and rather engineer the new breed of applications to support offline modes of working. In order to avoid the need for manual deployment of apps on desktop (and associated management overhead), components enabling offline work would need to be downloaded and launched via browser. Although some of the foundational technologies that may be used for this e.g. (ActiveX, Java) have been available, i have not seen many applications built in this style. Without going into the underlying technical details, this would imply to me that this is not trivial.

Considering where we are with connectivity and state of art of web apps development, and considering how long it took us to get here, I'd guess that the challenges outlined above will take more than few years to overcome.

Having said that, there are probably plenty of opportunities to pursue the vision for less challenging parts of the workforce. Also, in the meantime, there are bound to be some more tactical solutions (e.g. VM) that could be used to make people lives easier.

I think we're on the same page here, as I think that we're looking at a multi-year journey with some very challenging problems on solve on they way, and we do need to be looking at some tactical solution in the mid term (such as VMs); however I do get a strong sense that it's a journey worth taking as the potential payoff is huge.

Web 2.0, thin, thick client etc. all will have their place. Lets look at another trend the move to laptops ie. the biggest selling pc presently is the laptop and indeed the percentage of laptops to desktops in a lot of our customers is moving towards laptops. This is indicative of how people feel about their workspace ie. why they do not like hot desking, why CRTs and LCDs are festooned with small creatures and photos of the kids. People like to own and personalise their space and be "in control" of the environment. VDI or PC blades providing the desktop experience on thin client devices is one of the possible ways forward. With connectivity becoming more available and reliable we may see Compaqs thin client mobile device reappear etc. etc. The simple matter of fact is that we will see more diversity in approach not the death of one or other option! (look at the possibilities - thin client with a thick client option if required eg. Soft grid, just thin client citrix/terminal services, the pc as is, thin client accessing thick client back end (perhaps a pc or server blade with XP) where a pc environment is provided in a virtual environment using rapid provisioning methods(something Capgemini has been doing since server blades and VMware first appeared), etc. etc.

Do you know of an online collaboration project management that can be installed on your site? Seems like it would be a big SEO boost to keep it on your domain name.

Thanks.