« How not to be outsourced (and keep your job) | Main | Is your SOAP turning to SOUP? »

There is so much excitement at the moment over applications delivered via a Web browser - wikis, blogs, shared documents/workspaces, portals, Web-based messaging, and so on - that no-one seems to be asking how much any of this actually contributes to the quality of workplace software.

It's all very new, which is an attraction in itself.  And it appears easy to start using such applications, since their user interfaces are often simple.  But is this type of software application actually any better than we had before?

In fact, it's usually much worse.  We've got used to complaining about having too much functionality in applications such as Microsoft Office, so at the moment the Web 2.0 stuff is benefiting from a rebound effect - it's simplicity is tempting. But here are some reasons why such software is a poor substitute for old-fashioned desktop software:

• It is not sensible to try and provide the same level of user interface functionality in a browser application that you can in a desktop application.  Browser programming technology is too lightweight to support truly sophisticated client applications - Javascript, to take one example, is a simple interpreted scripting language that was never intended to be as robust or powerful as the Java language it mimicked.  Any attempt to provide enterprise-class user interfaces solely via such browser technologies will ultimately founder, due not only to problems of download time for huge applications, but also because it is building a house on sand.  Using Web 2.0 means accepting a simplistic user interface.
   
• Most organizations are drowning in complex middleware systems, each of which carries a very high total cost of ownership. No-one wants more such systems. Desktop applications tends to be lightweight - the barrier to adoption is low, whether for one person or for the entire organization, since there is no need to involve the IT department at all. Web 2.0 applications, however, are the opposite - to the user, it seems simple to get started, but most commercial companies do not want their wikis or documents hosted on Google. They want them on their own servers, accessed by server-side software that they have installed, which means engaging with the IT department to start yet another long and involved software adoption cycle.
  
  
• Browser applications intended to support collaboration, as many do, depend on all concerned having access to the same servers.  But collaborative human work processes typically involve people from more than one organization.  Hence all participants may not have access to the same servers, or to the necessary resources on those servers - and even if this could be provided, there are many situations in which no one organization has the right to "own the process".  Hence it is necessary for all concerned to work in peer-to-peer fashion, rather than via a centralized server bottleneck.
   
• Systems intended for use by all kinds of people in all kinds of situation must be very easy to use, doing most of the work for you invisibly - especially back-end integration such as data sharing between applications.  This is not at all true of "Web 2.0" applications - they all present a different user interface, all run on different servers, and do not interoperate in any consistent fashion.  However much office workers may complain about Microsoft Office, they have come to expect a single user interface, which intelligently manages data integration complexity on their behalf.
   
• Any Web-based application is available only when you are connected to the Internet.  But in many working situations no such connection is available - when you are in transit, for example.  A well-designed desktop application, however, can be used at any time, even on any device.  Put the program and your data files on a USB stick, for instance, and you can work wherever you are, using any computer to which you happen to have handy, whether or not an Internet connection is available.

The most successful new desktop software applications of recent years - Groove, Skype, Firefox, and Thunderbird, for example - have not been browser based.  And if they were built today, using currently popular AJAX technologies, they wouldn't work as well as they do - they wouldn't be as fast, as stable, as well-featured, or as easy to use.

So when you are next considering adoption of an end-user workplace application - or building one - ask yourself this question:

Will a browser platform provide the level of functionality that, over the 20 years since the emergence of windowed operating systems, people have come to expect?

And don't be surprised if the answer is a resounding No.

Posted by keithhb in Internet • Office Applications |Digg This|Add to del.icio.us