Transitive’s software portability promise
- Article 8 of 25
- Infoconomist, August 2001
Transitive claims that its software allows users to 'write once, run anywhere', but whether potential customers will listen is another matter.
There are certain orthodoxies that every developer quickly learns: hardware solutions are faster than software ones; emulation is slower than native code; and programs written for one chip will never run on another.
But Transitive Technologies claims that it can overturn such entrenched conventions. It says that its Dynamite software not only speeds up existing programs on their native, pre-ordained platforms, but also on completely different ones, translating them on the fly.
Put simply, Dynamite takes code that is intended for one processor, before passing it onto a translation 'kernel', and then moving it on to its target processor. “The Dynamite kernel is production-ready. The key to our technology is that we are able to get code into an intermediate format very, very quickly. Most translations take such a long time they can't do many optimisations,” says CEO John Graham. The company has already demonstrated software intended for standard Intel x86 processors running on MIPS-based servers.
Development of the target processor back-end can take as little as four to five months. However, some parts of the process are not so easy and require pre-written libraries of programs. “You don't take it out of the shrink wrap and expect it to run. It's not that kind of silver bullet,” says Graham. “Dynamite is a partnering and licensing proposition.”
So far, forays into binary translation technology, such as Transmeta's Crusoe chip, have met with indifference. Nevertheless, Transitive has backing of $3.3 million (€3.7m) from Manchester University's investment fund and Pond Ventures, and it will be looking for a second round of funding in 2002.
In claiming that it is able to run any program anywhere, with “1:1 or even better than native speeds”, the company has given itself much to prove. Indeed, such claims as those once made by Sun Microsystems about its Java programming language – “write once, run anywhere” – are now widely derided.
The company therefore needs to convince people that its technology is more than a technical sleight-of-hand. The software needs to work well, and, crucially, it needs to show that it can indeed bring speed benefits and portability to platforms and products that were never meant to be translated to other environments.