Something interesting is happening in the application development world: the Web browser is making a comeback as a means of delivering standard business applications to users.
Despite their many disadvantages over locally installed programs, browsers do just fine at many standard applications such as reading e-mail, managing entire enterprises (ask SAP), controlling remote programs and servers, and of course browsing online services like the Web.
This browser-is-the-platform prediction has been made before. Ten years ago, when Netscape was in full public offering cry, Microsoft executives were rightly worried that the browser would become the standard application delivery platform. They made sure that their own browser squeezed Netscape's off the Windows desktop. This strategy got the company convicted for antitrust violations under the US Sherman Act in the US.
So what's different this time? Two things, both of which have far-reaching implications - read opportunities - for local software developers.
Firstly, Microsoft has managed to alienate many application developers over the past couple of years with too much chopping and changing of development methodologies, and - a potentially disastrous move - by breaking backward compatibility in some of its most popular tools. And there is the promise of more breaks to come in future versions of Windows.
Developing applications for Windows, the largest software market available, requires learning the Windows application programming interface (API), the hooks into the internals of Windows that Microsoft makes available to developers so that they can write applications. Microsoft has traditionally worked hard at keeping its APIs backward compatible so that existing code can be moved to new versions of Windows with as little trouble as possible. More developers with more experience of Windows means more applications for Windows - and applications are what drive operating system sales.
But that future compatibility is now in doubt and developers are responding accordingly. Faced with two or three years of potentially wasted effort in programming for a technological dead end, many developers have either moved to Microsoft's existing Web-based tools or moved to other platforms.
A respected software development commentator and ex-Microsoft employee has recently claimed that this trend means Microsoft has "lost the API war". The Web markup language HTML is what will matter in the future, he says - not Windows. Windows clients will still be around for some applications, but mainly for running Web browsers. And Web browsers are now very smart indeed.
The second difference between 2004 and 1994 is that modern browsers have the capability to deliver rich applications to users - no matter where they are in the world or what devices they use to access the network.
If these two trends continue as they are, the local software industry has plenty to gain. Local developers are more than competent, and have been for many years, but have had to compete against packaged application software from overseas vendors. If the world moves to browser-based applications, local could become lekker very quickly.
Developing applications that run on a central server is cheaper and more friendly to developers, which means cheaper, more flexible and more effective for local customers. Instead of an upgrade cycle of six months to a year, server-based software is easy to improve continuously and with upgrades that don't tie up the corporate desktop fleet for days at a time.
It is still perfectly possible for a small team of programmers - perhaps just two or three - to produce something compelling and useful just by working hard and smart from the garage. Watch out for such a team coming to fulfil a need somewhere near you soon.
- Paul Furber is a freelance writer and developer of continuously improving server-based applications
