Loose Ends by Sydney Brenner (reference to use of TRAC language)
1
*Io~e J am a compulsive bookshop browser 1Jo,05 l and have spent many happy hours e^%d S learning all manner of things from hnnlcc T rnnirl npver affnrd t n nx.m I particularly like reference books - dictionaries, etymologies and collections of quotations and euphemisms. It is in bookshops that I have discovered that nearly all my ingenious theories of word origins are totally wrong, and I am now planning to be the would-be author of a should-be etymology. Tr At-Pr riav .7hlP n-n1h-lAn l-- · 1~ II 'V~ll Utl.l Udy, WIILC llvllbtCIlldcllLly paging through the index of a rather ~ O ~0['14,~ expensive new book on humorous 87tf,'P-e e*- quotations, I was pleasantly surprised to find my name. Hastily I turned to the page indicated. Was it 'He was a man of slender ends' or, perhaps 'To go tongue in hand'? No. It was some remark I had made about computers being poised between the obsolete and the non-existent. Actually, I made this remark about buying a computer. It is now nearly thirty years ago that I found myself trying to acquire a research computer to help us in the recon- struction of the nematode nervous system. I discovered that only two kinds of machine were available. There was one I could have immediately, but it was obsolete, small and slow - really a pile ofjunk. The other was a wonderful machine, bigger and faster with a radically new design, which was actually being built at that very moment and which I could have in six months time. I realized that this choice between the obsolete and the non-existent would always be there, because in six months or a year's time, the same would be true; last year's non- existent machine would materialize into instant obsolescence and there would be a better one to wait for. My conclusion was that if we followed this logic, we would never buy a computer. Instead, we bought one which had just appeared and occupied a metastable state between reality and the dream. It was a marvellous British invention and the exact opposite of user-friendly. It found all users most distasteful and seemed to go out of its way to avoid any constructive interaction with us. Most of the time it responded by going absolutely silent; at other times it issued cryptic remarks before collapsing completely. As was typical of such advanced machines at the time, there was very little software; there was only an assembler and programs were written and punched on paper tape for entry into the machine. Our paper tape reader sometimes had outbursts of hysterical rage and the tape would emerge neatly shredded into several strands. We wondered whether the reader had been developed under contract to some intelligence organization and we had unluckily been delivered one of the top secret models. As I am wont to repeat to my younger colleagues, this was the real way to learn about computing. I had found a description of a string processing language called TRAC in a journal and I decided to write an interpreter for the machine we had purchased. I reasoned that I could then free myself from the bonds of assembly language and write all my programs in the TRAC language. I managed to do this and indeed the first program ever to compare nucleic acid sequences was written in TRAC. It was used to look first at tRNAs and later at longer sequences. The most difficult part of the program was that needed to print the comparisons in panels on a teletype machine. John White was my teacher and he and I, together with David Marr, wrote a large amount of system software for this machine. I had become so skilled at assembly language programming that I didn't think twice about altering a Fortran compiler to use with our disc operating system. We also invented and built several gadgets, such as a cheap digitizing plate and a wonderful machine to handle electron micrograph plates for comparisons. All of this work became valueless when the machine - by then slow, small and obsolete - was junked and its pieces given to other people who could use them. My TRAC became useless, as the interpreter had gone with its unique host. I resolved never to become so involved with computers again, but I knew they were going to be essential tools in biology. Growing up with crystallographers really set the stage for our computing interests, which went beyond sheer number crunching. But most biologists thought all of this was a waste of time and all biochemists were convinced that computing was just an excuse to get off doing work at the bench. I spent many hours persuading people that computing was not only going to be the essential tool for biological research but would also provide models for analyzing complexity. The development of sequencing techniques and their widespread application has generated enormous databases of information, and the need for computers is no longer questioned. The amazing progress made in computers themselves and software for them, means that very power- ful models are widely available at low cost. There is no inflation for computer expenditure; one spends the same amount each year but every two years or so one gets three times more for it. With the development of computer networks, everybody will be in permanent communication with everybody else, all playing in the same gigaband. All except me. Some years ago, when I took up computing again, I decided to do better than the first time around. So I learnt the C language and wrote an interpreter for TRAC in C. I have a large suite of programs written in an even flashier TRAC language that I use to study sequences. I can run these programs on any machine that has a C compiler that can compile the interpreter. I have my private language that I doubt anybody else will want to learn. But I am no longer trapped in one machine and I can work happily with my Mac or with the most powerful, multi-this and multi-that supercomputer available. © Current Biology 1995, Vol 5 No 11 1328
Transcript of Loose Ends by Sydney Brenner (reference to use of TRAC language)
*Io~e J am a compulsive bookshop browser1Jo,05 l and have spent many happy hours
e^%d S learning all manner of things fromhnnlcc T rnnirl npver affnrd t n nx.m I
particularly like reference books -dictionaries, etymologies andcollections of quotations andeuphemisms. It is in bookshops that Ihave discovered that nearly all myingenious theories of word origins aretotally wrong, and I am now planningto be the would-be author of ashould-be etymology.
Tr At-Pr riav .7hlP n-n1h-lAn l--· 1~ II 'V~ll Utl.l Udy, WIILC llvllbtCIlldcllLly
paging through the index of a rather~ O ~0['14,~ expensive new book on humorous
87tf,'P-e e*- quotations, I was pleasantly surprisedto find my name. Hastily I turned to
the page indicated. Was it 'He was a man of slender ends'or, perhaps 'To go tongue in hand'? No. It was someremark I had made about computers being poisedbetween the obsolete and the non-existent.
Actually, I made this remark about buying a computer. Itis now nearly thirty years ago that I found myself trying toacquire a research computer to help us in the recon-struction of the nematode nervous system. I discoveredthat only two kinds of machine were available. There wasone I could have immediately, but it was obsolete, smalland slow - really a pile ofjunk. The other was awonderful machine, bigger and faster with a radically newdesign, which was actually being built at that verymoment and which I could have in six months time.
I realized that this choice between the obsolete and thenon-existent would always be there, because in six monthsor a year's time, the same would be true; last year's non-existent machine would materialize into instantobsolescence and there would be a better one to wait for.My conclusion was that if we followed this logic, wewould never buy a computer. Instead, we bought onewhich had just appeared and occupied a metastable statebetween reality and the dream. It was a marvellous Britishinvention and the exact opposite of user-friendly. It foundall users most distasteful and seemed to go out of its wayto avoid any constructive interaction with us. Most of thetime it responded by going absolutely silent; at other timesit issued cryptic remarks before collapsing completely. Aswas typical of such advanced machines at the time, therewas very little software; there was only an assembler andprograms were written and punched on paper tape forentry into the machine. Our paper tape reader sometimeshad outbursts of hysterical rage and the tape wouldemerge neatly shredded into several strands. We wonderedwhether the reader had been developed under contract tosome intelligence organization and we had unluckily beendelivered one of the top secret models.
As I am wont to repeat to my younger colleagues, this wasthe real way to learn about computing. I had found a
description of a string processing language called TRACin a journal and I decided to write an interpreter for themachine we had purchased. I reasoned that I could thenfree myself from the bonds of assembly language and writeall my programs in the TRAC language. I managed to dothis and indeed the first program ever to compare nucleicacid sequences was written in TRAC. It was used to lookfirst at tRNAs and later at longer sequences. The mostdifficult part of the program was that needed to print thecomparisons in panels on a teletype machine.
John White was my teacher and he and I, together withDavid Marr, wrote a large amount of system software forthis machine. I had become so skilled at assembly languageprogramming that I didn't think twice about altering aFortran compiler to use with our disc operating system.We also invented and built several gadgets, such as a cheapdigitizing plate and a wonderful machine to handleelectron micrograph plates for comparisons. All of thiswork became valueless when the machine - by thenslow, small and obsolete - was junked and its pieces givento other people who could use them. My TRAC becameuseless, as the interpreter had gone with its unique host. Iresolved never to become so involved with computersagain, but I knew they were going to be essential tools inbiology.
Growing up with crystallographers really set the stage forour computing interests, which went beyond sheernumber crunching. But most biologists thought all of thiswas a waste of time and all biochemists were convincedthat computing was just an excuse to get off doing workat the bench. I spent many hours persuading people thatcomputing was not only going to be the essential tool forbiological research but would also provide models foranalyzing complexity.
The development of sequencing techniques and theirwidespread application has generated enormous databasesof information, and the need for computers is no longerquestioned. The amazing progress made in computersthemselves and software for them, means that very power-ful models are widely available at low cost. There is noinflation for computer expenditure; one spends the sameamount each year but every two years or so one gets threetimes more for it. With the development of computernetworks, everybody will be in permanent communicationwith everybody else, all playing in the same gigaband.
All except me. Some years ago, when I took upcomputing again, I decided to do better than the first timearound. So I learnt the C language and wrote aninterpreter for TRAC in C. I have a large suite ofprograms written in an even flashier TRAC language thatI use to study sequences. I can run these programs on anymachine that has a C compiler that can compile theinterpreter. I have my private language that I doubtanybody else will want to learn. But I am no longertrapped in one machine and I can work happily with myMac or with the most powerful, multi-this and multi-thatsupercomputer available.
© Current Biology 1995, Vol 5 No 111328
