High-tech

Java - drowning in alphabet soup?

Java was heralded as a panacea for developers and users, but with many different flavours emerging The Chilli dips into the 'soup' to see if there are still opportunities for the entrepreneur in this field.

When the concept of the Java language and virtual machine environment was introduced, it promised application developers the panacea of "write once, run anywhere". The dream of writing an application once and then being able to run it on any platform or machine, without having to edit, recompile, or retarget, was meant to develop a whole new industry sub-sector. It was meant to be the lifeblood for many new start-ups.

Fast-forward a few years and the reality has turned out to be different. Many entrepreneurs, start-ups, investors, service providers and terminal vendors in the embedded space are still in denial, not because they are blind to the truth but because after spending millions developing a whole new industry, one has to soldier on, until either the cash runs out, or somebody finally figures out that they have been taken for a ride.

Java may have been reasonably successful in the enterprise space, where there is plenty of cheap processing power and storage (PCs, workstations) and the number of varying platforms are limited to Wintel and Unix. When it comes to the embedded space, for example mobile phones or set-top boxes, or even the humble SIM card, Java has simply failed to deliver on its promise.

When Sun agreed to act as the master co-coordinator for the many incarnations of Java, it had underestimated the effort and resources required to unify the many warring factions in the embedded space - a task normally reserved for standard setting organisations such as ETSI or ANSI rather than a single commercial entity. Most of the players in the embedded space have conflicting agendas and have not only carved out their own little Java beans but have started hoarding some of them.

So here we are at the end of 2002, and we are sill waiting for the "write once, run anywhere" dream to turn into reality, with much of the functionality, such as user interfaces and security, still not resolved or awaiting some committee decisions.

Mobile network operators (MNOs) have been blaming the handset guys for not delivering interoperable 2.5G or 3G phones. The handset guys have been blaming the many unresolved issues such as memory footprint, processing performance and power issues and to add to that, there is the thorny issue of who is entitled to the royalties, and how much?

When Java was first widely promoted, there was a whole plethora of new start-ups backed by some seasoned angel investors and entrepreneurs, who at the time relished the idea of being the next Intel, which optimised the x86 to run Microsoft Windows. The thinking of the investor community was along the lines of "If only a few of these start-ups could optimise their Java on a processor design and file a few patents and hold on until their competition weathered away, we could be backing a potential new Intel". The question that the investors and start-ups forgot to contemplate was which flavour of Java they were going to back. MIDP (Mobile Information Device Profile), J2ME (Java 2 Micro Edition), Java Card (for smart cards)? The goals were never set, boundaries were never delineated and each variant has evolved at an asthmatic snails pace, such that companies with limited resources and under time pressure have created their own variants.

Several MNOs took this route. The press has been full of praise for NTT DoCoMo, who has generated substantial data revenues with their I-mode service, but DoCoMo has had to create their own proprietary version of Java MIDP 1.0, DoJA (DoCoMo JAVA), with its own i-appli profile. This was necessary as DoCoMo wanted to encourage a whole community of independent application developers that can rely on a stable DoJA platform.

So if you happen to be a Vodafone subscriber from the UK on a trip to Japan and want to download an application to your Nokia handset, will you able to do it? You have guessed it - no! The developer will have to re-write it for VoJa (The Chilli shorthand for Vodafone's Java platform). On the other hand, should the cash-strapped developer write for NoJa (Nokia's Java), or would it be more profitable to target MoJa (Motorola's Java), or shall we have SeJA (Sony-Ericsson Java) or SaJA (Samsung JAVA)? On top of the handset variations can be added the different operator variations, e.g. OrJa (Orange's Java), TeJA (T-Mobile's Java) or TikiJa (Java for Tikiphone, an MNO in French Polynesia).

The current clutch of Java-enabled handsets is slow to respond to user commands (speed), take a lot of memory (cost) and consume too much power (battery life). Add to that the different flavours of Java, and the door is wide open for an alternative solution, which only the inventor of Java can close, if it is really willing to go head-on with its archrival Microsoft. Microsoft hasn't been resting on its laurels. Microsoft has been busy working directly with most major operators such as Vodafone, T-Mobile, Telefonica, Orange, Singtel, ATT Wireless, Verizon and Cingular. That is a steady base from which to slowly throttle Symbian and convert the handset folks over to an integrated Microsoft offering and this may not include Java but their own version of wireless .NET and smart card .NET or simply plain old C, C++.

For the entrepreneur, this could be a two-edge sword; it could eliminate many of the hardware based Java companies but at the same time, create opportunities in application and tool development. On the other hand, if your team's skills set isn't that portable, then it may be advisable to start considering some other alternatives, as ants do not survive well when bears attack their nests.

Future articles in The Chilli will examine these new growth areas, where ants can beaver away and still collect enough food to survive the current downturn.

Comments on this story? Send an e-mail to picasso@theChilli.com.

© Chilli Publishing Ltd 2003

18DEC2002