SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Stock Swap

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Trader X who wrote (10964)1/29/1998 3:22:00 PM
From: Gene Veinotte  Read Replies (1) of 17305
 


Software: Learning to talk with genes
TUESDAY DECEMBER 23 1997
--------------------------------------------------------------------------------
Mark Ward on an inspired approach to writing and running software
--------------------------------------------------------------------------------
A 60-year-old thought experiment has inspired a novel approach to writing and running software that could one day make operating systems and memory intensive applications obsolete.

In 1935 Cambridge undergraduate Alan Turing - the mathematical genius who later helped to crack the German Enigma code - came up with the idea for a universal computer that could perform the functions of any other, more specialised, computer.

This device worked by scanning a tape fed into it and reacting on the basis of the symbols printed on the tape. Mr Turing never intended the machine to be built, but worked on it because he was interested in what problems were and were not computable.

Now Bernard Hodson of Genetix Software in Ottawa, Canada has taken Mr Turing's ideas and used them to inspire an approach to software writing that could lead to smaller, more robust computer programs and faster application development.

Existing computer programs act like arrogant snobs - they do nearly everything their own way, rarely sharing computer code even if they are doing the same thing (such as displaying text on screen) as another program. Some Microsoft and Apple Macintosh programs do call on the same bit of computer code for some functions, on PCs and Macs respectively, but only for a fraction of what they are capable of carrying out.

In contrast, Genetix software programs share everything. Unlike Mr Turing's theoretical machine their instructions are not written on a tape, they are contained in what Mr Hodson has dubbed "genes." The most basic genes are written in machine code and when they are called on to perform they talk directly to the processor in the computer. This saves space and removes the need for an operating system.

Higher functions are built out of lower level genes. Creating an application is relatively straightforward. The genes that already exist can be used for most of the functions, but some new ones may have to be written for any functions unique to a particular application.

"We take an application and see what genes it needs," says Mr Hodson, "Then we build the genes if there is something that we have not previously written." He says this is one of the main advantages of his approach: the more programs written using Genetix software the bigger the pool of genes and the fewer new genes are needed to emulate other programs.

One thing Genetix software does share with Mr Turing's machine is a single reading head that works out what to do. This tiny program calls up the genes from a common pool that do what the user of the program wants to.

Using this approach Mr Hodson believes it would be possible to do many things larger programs such as Microsoft Word do, but in far less memory, possibly only a few hundred kilobytes.

For the moment though, Mr Hodson is concentrating on more specialist applications. He is working on using the Genetix approach to make the processors on smart cards more powerful and to send video down telephone lines.

Next year he hopes to produce a version of Genetix that can run programs written in Sun's Java computer language. He is also planning a European conference next year to present the latest work.

Robin Bloor, a UK-based computer consultant, says he is impressed so far by what he has seen of Mr Hodson's work. He believes the Genetix approach will find its first uses in smart cards, embedded processors and for programs currently being written in Java.

Mr Hodson, a retired computer professional, says he has been thinking about the ideas for the Genetix software for years, but only now does he have the spare time to develop them. Although he started later than Mr Turing his legacy could be just as long-lasting.

Mark Ward is technology correspondent of New Scientist.

--------------------------------------------------------------------------------

For other articles on this subject, try Site Search and The Archive

--------------------------------------------------------------------------------






c Copyright the Financial Times Limited 1998
"FT" and "Financial Times" are trademarks of The Financial Times Limited.
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext