Paradigm Shift in Computing
Million of years ago, the salmon were fish who lived in streams and bred in ponds in the streams. Male and female salmon found each other by synchronising the location using specific chemicals present in the water, and synchronising the time using the position of the moon. Through this protocol, the salmon would feed where they pleased, but they would come back to the pond, year after year, in order to breed.
Over the years, the path of the streams has changed, the continents have drifted, the salmon have taken to the sea, but one feature has not changed: they return to the same ponds for breeding. They sometimes travel thousands of kilometres, and jump their way up waterfalls, in trying to reach the designated pond.
Why do they persist in doing this? Would it not be optimal for both males and females to find some other way? The problem lies with the "industry standard" that has been adopted. If one individual fish chooses to violate the standard in isolation, it will not find a mate and perish.
The predicament of the salmon powerfully illustrates the power of "industry standards" to enhance economic efficiency by enabling cooperation (the ability of the salmon to find each other), or destroy economic efficiency (the irrationality of swimming thousands of kilometres to the designated pond at the designated time).
All PC users have known the frustration of the "upgrade cycle". Most spreadsheets a few years ago worked on a 386 with 8M. Today, in order to do the same work, users need a 586 with 32M. The work has changed little, but the computer resources required have changed a lot. In addition, corporate IT departments are going through huge costs every year for the upkeep of these machines: installing software, reducing the frequency of crashes, removing viruses, resolving conflicts between hardware, etc.
Is there something going wrong here? Year after year, IT budgets invest in enhancing desktops, but the functionality of the desktops stays mostly unchanged. The great leaps of information technology applied to business have taken place through large corporate computers and communications; not through improved features in the word processor used at desktops. Indeed, the great advances in hardware should have made desktop computers much cheaper; however, the bloating of software has kept hardware costs at the desktop roughly constant.
There is a very real sense in which the worldwide PC industry is locked in the same problem as the salmon. There is a large installed base of users who have certain hardware and software (PCs with Microsoft Windows). Software developers write products for this, despite the well--known engineering deficiencies of PCs and Microsoft Windows. This continues to attract users to this class of hardware and software. The desire for compatibility keeps this vicious cycle going. In economic parlance, this is a situation with a network externality, driven in part by the peculiar cost structure of computer software, with high fixed cost and near--zero costs of copying.
There was a famous episode two years ago where typing at 40 words per minute was not possible using Microsoft Word for the Apple Macintosh running at speeds below 40 Mhz (that was when visions of salmon jumping up waterfalls first came to me). I have known entire bank branches to be computerised using the same CPU as the Apple Macintosh, running at 40 Mhz. It was a new low for economic efficiency when this computer power became inadequate for a single user!
The way out of this situation has come about in an unexpected manner. The first step was taken by the explosive success of the Internet, the twenty year old network built by universities in the US. The internet, and the web browser, proliferated to desktops all over the world in 1996. Today, the web browser is on every desktop in the world. Together with a new technology called Java, the web browser can be used to run software.
This is a moment of enormous significance. Traditionally, software developers wrote software for Microsoft Windows, because most users had Microsoft Windows. The availability of software for Microsoft Windows forced users to continue buying Microsoft Windows. Today, software developers can write software for the new standard, which is the web browser coupled with Java. Software developers benefit from this, because these new technologies have grown without the distortions of the Microsoft approach. Software developers can do this without fear, because web browsers are ubiquitous. Even PCs with Microsoft Windows are hospitable to such software, as long a web browser is installed.
This is a breakthrough of enormous significance for the computer industry. It frees the salmon from the need to visit the same pond: computer users and software developers can now proceed on finding economic efficiency for each other, free of the dependance on PCs running Microsoft software. A simple example of the gains in economic efficiency is the "network computer" or "java terminal", a device which costs under Rs.25,000, plugs into the corporate network, requires zero maintenance or upgrades, and runs software written in terms of the web browser and Java. This device generates enormous savings for companies. Major companies like Nomura and Federal Express have initated programs to replace tens of thousands of PCs using java terminals.
These developments are unpleasant for Microsoft, a company which has enjoyed a monopoly status for nearly two decades. It has now been presented with a new competitive territory, where it has no special advantage as compared with nimble, young rivals like Netscape.
The strategy that Microsoft has now adopted is to try to win back some of its former role by making a Microsoft--version of the Internet. (This is what has led to the recent lawsuit about Java). If users try to connect up with software developers through Java and web browsers, but the Java and the web browser are modified so that they only work under Microsoft Windows, then the dependance on Microsoft continues. Microsoft has taken the highly unusual strategy of giving away its web browser for free in an attempt to make these Microsoft--versions popular.
The strategy of making non--conformant Java is a fairly dangerous one for Microsoft, because it market share in the browser business is only 30%. Of the 116 companies which have licensed Java from Sun, Microsoft is the only one which plans an incompatible variant. Microsoft hence faces the acute lack of acceptance which comes from not using a major industry standard.
What does all this mean for the corporation? How does it avoid the predicament of the salmon? The road ahead for companies is based on three principles:
- Stop expenditures for the `upgrade cycle' on PCs,
- Target the web browser and Java in software development,
- Avoid software from Microsoft in connection with the Internet.
These guidelines generate cost savings for companies today, and leave the window of opportunity for more radical cost savings later on. An existing PC running Microsoft Windows can be made useful immediately for the new generation of software by installing a web browser and Java. Once this software is proven, the transition into network computers, with further cost savings, can come later.
Back up to Ajay Shah's media page