Ramine
7/6/2014 12:47:00 AM
Hello,
You have seen me talking about the advantages of high level
programming languages over low level programming languages...
I have said that since high level programming languages
do bring bigger and safer and even faster constructs so they enhance the
criteria of reusability and since they enhance the criterias
of reusability so they enhance the criteria of productivity...
Other than that since high level programming languages bring
high level constructs those high level constructs can be
made fast and safe , so high level programming languages
can enhance the criterias of reusability , productivity and
performance and safety...
And since we are in the era of parallel programming, we have
to think better , cause i think that high level programming
languages do there best to easy for us parallel programming,
so think twice before using low level programming, cause
low level programming is error prone and difficult and is
bad for parallel programming, cause low level programming
misses high level constructs that are much less error prone
and misses high level constructs that easy the job of parallel
programming... so as you can see the advantages of high level
programming languages are becomming more and more clear and more
important and i think they outnumber the advantages of low level
programming.
Another advantage of high level programming languages is the following:
When you take a look at the derivative in mathematics, the formula
of the derivative of 2*X or a*X that is a, since we know that
lim x2-x1 -> 0 de (2*x2 - 2*x1)/(x2-x1) , so we have that
2*(x2-x1)/(x2-x1) , so it's equal to 2
So as you have noticed i am using the "Limit" and some mathematical
calculations to be able to proove and to generalize and to find the
derivative of a*x, those mathematical calculations that we must do
to proove the general formulas of the derivatives are low level like
assembler and they allow us to know from where the general formulas of
the derivative of a*X comes from... but as you can see there is also the
short way of learning is to know that the derivative of a*X is equal to
"a" without proving it, and this way of learning do enhance the criteria
of reusability and productivity also and do permit us to construct other
ways of learning, so there is mathematicians that must learn how to
proove the theorems and the general formulas and there is others
engineers and like that have to learn only how to work in a high level
ways with mathematics without prooving theorems or general formulas.
That's the same for informatics.
As you have seen me talking in my previous post titled:
"Another advantage of high level programming languages..."
I have tried to explain to you that there is another advantage in
high level programming languages... and i have talked about the
other way of learning and have said that another way of learning
in mathematics is to try to minimize the number of proofs that
we have to do, so i have said that pure mathematicians have
to proof the theorems and general formulas, but others engineers
and like have only to learn how to work in a high level manner
like in high level languages without proving theorems and general
forumulas or keeping at the minimum the number of proofs that we have
to do, this way of learning do enhance the criteria of reusability
so it enhance the criteria of productivity i think...
As an example, you have seen me inventing some scalable algorithms
like my scalable RWLocks and my scalable MLock etc. so my way
of learning is high level and also low level since i have to understand
how to minimize the cache-coherence traffic and to minimize contention
and how to write some assembler code etc. to be able to invente my
scalable algorithms, this way of doing is honorable , but there is
also other high level ways of learning, like to just understand how
to minimize the cache-coherence traffic and how to minimize contention
and how to reuse my scalable algorithms without writing or inventing
those scalable algorithms, this high level way of learning is also a
good way of learning.
I have come to an interresting subject...
You have seen me talking about advantage of high level programming
and high level ways of learning...
But there is still an important desavantage of high level programming..
I explain..
You have seen me talking about the derivative of a*x or 2*x that is 2,
since:
Limit x2-x1 -> 0 of (2*x2 - 2*x1)/(x2-x1) , so we have that
2*(x2-x1)/(x2-x1) , so it's equal to 2
so follow with me please, the calculation above to proove that
the derivative of 2*x is equal to 2 or the derivative of a*x is equal
to "a" is a low level way of doing mathematics like in assembler
programming , we have to use the "Limit" and mathematical calculations
to proove that the general formula of a*x is equal to a.. but if you
look carefully, when we are proving and infering the general formula
we are like automating the derivative , since as soon as we proof
the general formula we don't have to proove again the general formula
and we have just to use the general formula, so by avoiding
to proove again the general formula we are like automating the
the derivative , it's like automation in robotics,
and we are speeding the process of calculating the derivative,
since we have not to do the proof again so in time t2-t1 of
1 hour we can do many more calculations of derivatives per hour,
so since we are not proving again the general formulas of derivative but
just using the general formulas we are speeding the rate at wich we can
do derivatives calculations per hour, and this apply to all mathematics
, since we are proving theorems and general forumulas and automating all
the mathematics, hence the rate at wich we can solve mathematical
problems will be faster, so we can do more mathematics per hour time and
this is the same for informatics, since we are constructing more high
level constructs that are safer and faster we are automating informatics
by enhancing the criteria of reusability
so by using those high level constructs the rate at wich we can
produce code and solve computing problems per hour will be faster, so
since the rate by wich we can produce code and solve compuyting problems
will be faster less and less programmers will be needed and less and
less mathematicians will be needed i think so this will higher the
percentage of unnemployement i think, and this is one important
desanvatage of high level programming i think.
Thank you,
Amine Moulay Ramdane.