output=function(input)

i have been mucking about learning new languages... i am sure they can be most eloquently expressed... but i am afraid my grasp of each of these computing languages is only enough to get by doing what i need... sometimes a little more... and often i am mentally challenged by having to decode one language and recode it in another because i am a little more familiar with the other language environment...

one of the oldest programming languages, FORTRAN, is still much loved by scientists and engineers of a particular generation... it has many merits in that, while archiac in its syntactic structure, it enables lightning-speed computation and hard-core number crunching... however, it takes a long time to get used to its cumbersome way of coding each step of an intent... other languages, e.g. MATLAB, works at a higher-abstract level, leaving the specification of each step of the mundane work e.g. reading each single unit of information from your file to in-built functions that you can simply use... being more intuitive, unfortunately, has its price... it takes longer to run a similar script in MATLAB than in FORTRAN... because the former 'normally' interprets rather than execute a compiled script...

but whatever the 'architectural' differences in the programming language... the most annoying is that different generations of researchers have their own preferred scripting tongue... and often because of differences and incomprehensibility (cryptically idiosyncratic) or lack of 'incentive' or opportunity (etc.) to acquaint oneself with the other... both group of users are often unable to freely communicate and share their spiffy codes... unless of course you spend time learning enough of one to decipher it... (well, more importantly i suppose, is to be less skeptical about the other language as a first step in breaking down the barriers...)

yesterday i spent the whole day translating about 2 pages of FORTRAN code into MATLAB so that i could benefit from implementing a code from my colleague to filter signal noise in my pilot-data without too much agony in reading my data into the FORTRAN environment... there is another script which enables the designing of a specific filter (Finite Impulse Filter) which goes with the former code but it is 6 times as long... and i am contemplating whether i should attempt to recode it... the pros could be that i might learn more about Fortran and have a better grasp of it, while improving my MATLAB skills by finding appropriate in-bulit functions for some parts of the commands... the cons are obviously time-related and errors... it is easy to mis-code a long script... sigh.

the biggest problem still... is to adequately filter out signal noise from the high sampling rates of over 1KHz for subsequent analyses... there are just simply many ways to tackle a problem... and it just takes a lot of trial and error to come to the best compromise between various parameters for a solution...

i never knew i would be dealing with so much programming... but i always never seem to know enough about anything...

posted by ~overacuppa~ on Friday, 24 February, 2006 at 23:29 hrs