Discussion Locked

I Just Want to Program! Don't Make Me Learn Math!

0 helpful
473 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

This is page 4 of the discussion. Read the opening post

473 replies

by kw08

I didn't attend formal education after 8th grade, so maybe I really missed something. I wouldn't hire a programmer that wasn't trained as a scientist. I don't remember the exact methods for Schroedenger wave equations, but I remembered learning it - I understand it. I hate accounting - I understand it.

Math? Shouldn't one be proficient in that BEFORE high school, by graduating HIGH SCHOOL, if you are going into a scientific DISCIPLINE, you better have calculus and the four laws down.

Gee, that second law comes up quick in life.

by tony hopkinson

Personally I think it's a good idea to hire trained programmers for programming jobs.

Most of the scientists I've met think they can program, evidence suggests they don't know crap about it. They are actually worse at it that engineers and business types.

We are talking about the people who take out option explicit from VB or use implicit in Fortran here.

I don't give a crap how good at calculus you are or which four laws you know, if you don't know it's a good idea to declare your variables and give them sensible names.

1st law of Coding is Murphy's Law.

by jean-simon.s.larochelle

...knows math very well:

f := a[i,i];
g := -sign(sqrt(s),f);
h := f * g - s;
a[i,i] := f - g;

This goes on for pages. This is a piece of Delphi code downloaded from the Web and I can tell you that this kind of code would not go very well through a code review by me.
A programmer should be a good programmer first and good at math second (or third because civilized might be second...or first?).
JS

by daveo2000

I have recently had my eyes opened in the process of interviewing for a new job. I have long lamented the lost art of efficiency in resource usage and stuck my tongue out at people who just throw hardware at a problem. I have now seen a place where it makes sense and, if we are lucky, it could have an impact.

In the banking industry, we are heavily dependent on computers. Everything needs to be done fast and so on. Most of the systems are, however, human limited in the sense that one of the links in the chain is a human. People are limited in how fast they can react to a given input so any such system will always be "slow".

The interesting application that I saw was for [b]programmed trading[/b]. Here, we have a case where a program is looking at data as it comes from the trading floor, checking price trends and, based on the very last trade, deciding to buy, sell or hold a stock (or other instrument).

When stocks were all traded by hand in the trading pits, the human was still an issue. Now, we have crossing networks where clients put their desired trades into the system and the system matches (or crosses) the trades. If the client is a person then their trade can be completed in microseconds and they will see the result a few seconds later when the screen refreshes. If the client is a program then the result will be seen in a couple of microseconds more.

With trading on this speed, and the fact that the big boys can all get their hands on the same kind of hardware, execution speed will, once again, come into the picture. In other words, the system with the most efficient methods and programs will "win".

We seem to have long since lost the focus on what computers are for (as in why do we want/"need" them). We focus far too much on what neat things we can do with computers (solutions in search of problems) and way too little on why we are doing it in the first place.

by joekarumba

It would be absurd to project the lack of English grammar mastery to over-indulgence in matters technical.
It is a fact that many IT pros have problems with languages.
I actually find it irritating reading posts with endless grammatical errors.
Actually, polishing the language skills is as important as learning new technical skills.
At the end of the day, communication skills supersede the capability to churn out marvelous code.
I think a course in Communication skills should be a must for CS students.
But away from that, we should not prejudge other people's English knowledge just because they made typo errors here.
That is insufficient evidence to prove deficiency of grasp.
And still you can ignore some petty lexical errors such as hazy, inconspicuous prepositions misfits.Especially when they originate from some of us whose primary language is not English!
Cheerz!

by tony hopkinson

However from the coding side of the job that should be writing readable code, which is mandatory for writing maintainable code.

Neither of these things are taught either.

Couldn't fit them in the course after the time allocated for calculus.

by jswalwell

I agree wholeheartedly that Math is essential for the programmer. I am an application developer for a division of a large company, and I use Math much of the time. There are many calculations that must be carried out in even mundane programs and I am often called on to create reports or views requiring complex calculations. Don't rely on others to write the formulae for you. Usually they have no idea how to get what they want and, if they do, they often get it wrong. Without Math I would have put out many innacurate business reports.

Also, I am appalled at the pathetic spelling and grammar seen in posts by those who are presumably well educated, for example, "its math" for "it's math" or "georges columns" for "George's columns" or "their" for "there". Programming is an exact science. It is unforgiving in the extreme. Minor typing errors can either cause the program to fail altogether or, even worse, run but produce innacurate information.

Having a good grounding in your native language is invaluable preparation for the precision of computer languages. Sloppiness in writing will be carried over to sloppiness in programming.

by absolutely

The keyboarding skills here exhibited by many -- not quite all -- participants are more consistent with the type of computer user known as "script kiddie" than with the type of computer user known as "IT professional".

Weird...

by tony hopkinson

polynomials, eigenvalues, asymptotic functions all essential to business reports.

Even basic arithmetic you must keep the axioms of commutation and association firmly in mind.

Natural language is incredibly useful. You can use your extensive vocabulary to come up with lots of meaningful names for the same function.

You can make you code richer and more expressive with alliteration and allegory.

You can lay it out so it matches the iambic pentameter, code as poetry.

Programming is an exact science? X-(

What planet are you from?

by daveo2000

Have you not seen the wonders
of writing good code?
How to overcome blunders
of a lost pointer's node?
There is only one way
for the cat to be skinned!
If you don't do it proper
you have certainly sinned!
Polynomials rule!
Eigenvalues, supreme!
Oh, the asymptotic functions
of which I have dreamed.
It's the math that is making
my heart go a-flutter
and the beauty of limits
keeps my mind from the gutter.
If you malloc some time
then you, too, will 'C'
that the way to Nirvana
and true freedom will be
in your deep understanding
of intercepts and slopes
which will better prepare you
to learn all the ropes.
After all, most the jobs
for programmers this minute
are for physics and science
and space travel, isn't it?
Visual Basic or Java or even Excel
aren't for real programmers,
anybody can tell!
The real programmers
are using assembly
and then calculate
with their slide rules, quite nimbly.
So, you see, my friend Tony,
how mistaken you are
there is only one way
(search you near or search far)
to solve any small problem,
big ones then even more-so,
'cause programming's exacting!
No room for a floor show.
Just because Perl provides
many ways to solutions
there's no reason to follow
such vague revolutions.
Your mind it too varied,
your thinking too scattered!
There's only one way to do it
unless you're mad-hattered!

by jswalwell

No one expects you to be Hemmingway, but the mistakes in your post are totally basic English. Nothing complex. I would never hire anyone who made so many mistakes because I would be terrified about what he/her would do in a program.

by jswalwell

I disagree entirely. A programmer lives by his keystrokes. If you can't type English (or whatever your native tongue is) correctly, how can you achieve any speed in programming?

I recall a story about Henry Royce, the engineer behind the perfection of Rolls-Royce. A machinist once took a part out of his lathe and commented, "There, that's good enough." He was fired on the spot. Royce's comment was that he couldn't tolerate anyone thinking that they could get away with less than their best.

by cettech

I agree with you that we should expect people to at least master the basics of English. However, in your post you made a few basic errors yourself. You misspelled the name of the author (it should be Hemingway, not Hemmingway). Your second sentence is not a sentence (requires noun + verb). Lastly your pronoun agreement in the last sentence should be he/she not he/her (one is a subjective pronoun the other an objective pronoun).

by jswalwell

My apologies to Mr. HeMingway. I was caught out by the pronunciation. My second sentence (Nothing complex.) was in the nature of an interjection, so I stand by it. It was more conversational than literary and may have benefitted by the addition of an exclamation mark. As for the he/her :-( ... What can I say? My face is bright red and I beg humbly for mercy!

by oarf

The study of mathematics is much more than add/subtract/multiply/divide. I would suggest that the appreciation of the correlation between the study of mathematics and application programming is one of the characteristics defining the difference between an average programmer and an exceptional programmer.

by alaniane

I disagree that programmers do not need higher level math. Calculus may not be used in a business application, but learning Calculus teaches you how to analyze problems. Programming requires that one fully analyze a problem. Personally, I think they need to make students learn the math without having the aid of a calculator. The math principles are more important than memorizing how to solve the equations.

by absolutely

Having seen both sides now run "get out the vote" campaigns -- one on MTV, the other door to door -- we see that just voting "D" or "R" without knowing something about the candidates, and ballot measures, is not sufficient to bring about the desired outcomes. Loyalists in both parties have very valid reasons to be disappointed, and not only with the opposition party.

by alaniane

Logical reasoning is an important aspect of mathematics. There does not need to be a one to one correlation for mathematics to be useful to programming. I would say that the mathematics that has proved most useful to me in programming has been Euclid Geometry. I do not know how it is taught currently, but when I was in school over 20 years ago, we had to write proofs for various theorems. I have yet had to code those proofs, but the concepts I learned from doing the proofs help me immensely in coding.

by alaniane

That is why I think they should ban the calculator from most mathematical courses. When you are forced to figure the problem out in your head instead of just punching keys, you start to see the value and beauty of mathematics. The problem I see with most mathematic and programming courses nowdays is that they teach you how to solve a problem by using a set formula. They fail to teach you how to derive the formula in the first place. Its the abstract concepts that are going to help you the most not the various formulas.

by tony hopkinson

Several branches of maths are useful models which can be programmed, ie applied.

I don't think anyone was arguing that maths isn't useful or that it doesn't have it's own Symmetry.

There are no axioms of programming.
There is no system of proof.
There is no one economic solution.

Using math to teach that lesson would be far more valuable, than attempts to treat programming as though the above myths were true.

I barely covered proofs in advanced math, (what I did I enjoyed).

Arithmetic well that's in another thread and you have my wholehearted agreement.

by alaniane

The point of proofs (at least when I was in school which is over 20 years ago) is to help a person learn how a solution is derived. Its actually a test of logic. Programming requires that I take a spec a break into smaller procedural blocks. Even if I am doing object-oriented programming, I still have to program procedurally aka methods/functions/subs.
When I prove a mathematical theorem, I have to show the steps I used to solve the problem and why each step is valid. Those are the principles that are important to learn. Its the logic behind the formulas and not the formulas themselves. That is the problem with many math classes today; people learn the formulas, but never see the basis for those formulas.

[Sorry I missed read your post about proofs. I concur that many of the math programs fail to teach sufficiently proofs. That is one of reasons I don't like having calculators used in a math class since teachers and students tend to focus more on using specific formulas to solve problems instead of showing how to derive those formulas.]

by oisleach9

If all you are going to do is program using Canned Routines, or perform basic Database operations then no advanced matematics is required. I can tell you that most of what I learned in attaining my MAster's in Chemical Engineering, or Applied MAthematics has never been used. But I also work with image processing, and have needed to use some of that knowledge to develop specialized routines. I have also needed my knowledge of statistics just to write and present decent reports on Sales database information.
My advice to anyone going into software development is learn statistics, matrices, vectors, and how to use them. The rest you can learn along the way. But be aware this type of atitude will limit you in the long run. How many candidates can go into a job interview and spout off about ANOVA, regression, trending, and time series analysis. This kind of info can put you over the next candidate.

by tony hopkinson

They want you to know transisolation, polymorphism, aggregation, referential integrity, transaction boundaries, coupling, cohesion, reference vs value, joins .....

I've used all the math you mentioned and more, never mentioned it in an interview once.
Image processing, signal processing, and high level stats are all niche skills.

They are obviously valuable where you've plied your trade, and you would whup my ass interview wise where they were valued.

None of them make you a better programmer though, just a more effective one in that area of the market.

Am I a 'better' programmer because I've done time in the manufacturing industry?

No math worth mentioning in that, but I helped save over ?100,000 pounds a year by the estimation of the manager who tasked me.
Simply knowing C & learning VMS saved a firm 80k a year in one job, I was there for six years.

You are talking problem domains.

by alaniane

I am not going to dogmatically state that Calculus is absolute necessity to programming. However, Calculus does hone one's analytical abilities. The problem I see with most people learning math (whether it is algebra, geometry, or calculus) is that they concentrate on the mechanics of the course. True, the mechanics of Algebra are used constantly through out life (not just in programming); however, it the principles behind the formulas that help us more. Knowing the logic behind differientiation and integration is more valuable than knowing that dy/dx x^2 is 2x.
The problem with the algebra courses that I took when I was in school is that they concentrated on teaching students how to solve equations by set formulas. If the student does not know the formula, then he is lost. Whereas a good geometry or calculus course tends to concentrate on the principles behind the mathematics.

I think both are essential to excel at programming.

by tony hopkinson

One of the most interesting things I ever did on a math course was learn the why of the formula for calculating the area of a triangle.

It was all done with pictures, by itself the math related to nothing I could picture, it was about as clear as a memory dump.

by alaniane

They really need to take the calculator out of high school math classes.
1) If you learn the basic principles, you can easily learn how to punch calculator keys later on.
2) In order to compensate for having a calculator, many math courses purposely complicate the problem. This causes the student to concentrate more on the mechanics of solving the problem and less on the principles behind solving the problem. If I have to use a calculator to calculate some incomprehensible irrational number, how am I going to see the basis for the calculation?

by alaniane

Whether you realize it or not, mathematics is a training in analysis and logic. If you are going to properly solve the problem domain, you are going to have to use some form of mathematics. How do you figure out which procedure to follow to accomplish the problem? It requires analyzing the needs of the customer and then finding the most efficient way to implement what they have in mind. You use the same concepts in solving mathematical problems.
You also learn how to deal with ambiguities in math.
example 3+5*2=? Is it 16 or 13? You have to define it before you can solve it. I have lost count of the number of times I have been handed some spec that was ambiguous. A good programmer is going to clarify what the spec actually means and not just code something and hope it flies.

by tony hopkinson

Programming requires that one fully analyse a problem ?

Not in practice it doesn't!

We can't even fully define the problem, how in Beelezebub's name are we meant to analyse it?

Programming != Mathematics

by alaniane

If you are going to define the problem, you will have to analyze it. You need to know what questions to ask to define the problem. How are you going to know the right questions? You may not explicitely analyze everything, but you better be able to implicitely do it.
I should retract "fully analyze" since in most cases it is impractical. What I actually meant is that you need to analyze to a practical degree the problem. Although, I frequently see code that hasn't been analyzed very well. A good example is not taking into account nulls or division by zero.

by tony hopkinson

You cannot mathematically analyse an undefined problem.
You can attempt to state the problem mathematically, potentially prove it, definitely disprove it.
Whether this math has anything to do with the real problem, which has not been defined is anybody's guess.

That's where you analogy of mathematical reasoning versus requirements analysis breaks down.

Both Null and div by zero are failures of representation, both can be dealt with by a richer definition of a 'number'. They aren't because it's impractical.

Both could be completely out of scope for the programmer.
If I was reading it as a type integer from a
Field in a table defined as
MyNumber int not null default 0

Should I check for null and div by zero ?

PS
By preference I'd check, and I'd do it in every routine that used the value as an integer and an extra check where it was a divisor.
That could cost big style though.

by deepsand

In those case where the systems analysis and the programming are separate functions, performed by different persons, you are correct.

However, it long ago became the practice of combining those functions, with both being performed by a Programmer/Analyst, regardless of whether or not that individual's title contained the word "analyst."

Thus, it is no longer possible to make broad assertions re. that knowledge & those skills which are unessential to the "programmer."

https://www.techrepublic.com.com/5208-6230-0.html?forumID=102&threadID=221757&messageID=2231694

by tony hopkinson

My job title has always had developer in it (well once I switched to IT), except when it was Systems Engineer.

I do systems analysis, never worked with one though, with me as the coder. Started in IT commercially in 87 as well.

Business Analysts / Domain Specialists, a few, very different job in my opinion.

I wasn't saying math was unessential, I'm dubious of pure math being essential.
I'm also made distinctly uneasy that there is an equivalence of concepts in mathematical reasoning and that done in programming or requirements analysis.

You end up with the Vienna Development method if you go down that route.

They are both systems of logic and to be successful, you have to express yourself correctly within that system.
Other than that they are wildly different though and that's before you get near implementation which is effectively what coding is.

by absolutely

Writing an app well means considering the resources available to it, and using them wisely. Fluency in mathematics will help, even if it isn't Absolutely Necessary.

by Former member

About three months ago I had to write a program that summed and unknown list of floats and had to branch a decision based on the error value. The program based its decision on an error value of 1^-13. call it fuzzy logic, if you will, or scientific computing.

Back in the day, on programmer that ended up in prison used the round off error in simple adding of floats to enrich himself. This was the banking system he bamboozled.

How much does accuracy in accounting mean to your company. It meant a lot to the banking system. The failure of an appropriate model also cost them 13 million before it was caught.

Based on this, I disagree that you can simple sum discrete columns of numbers.

by danlm

Everyone I have ever worked for has been completely happy with both my programming and communication skills.

1). My job is not to make you happy. Only the people that pay me. I'm still being paid and I am under no contract or any other binding agreement requiring them to keep me. They could hire anyone else at any time and I would have no legal recourse other then to find another job. In other words, I'm good at my job. Your loss, their gain.

2). I am not interviewing with you, again. I don't care what you think.

3). Communication skills are used so that others understand what you are presenting. I think I meet that requirement. Your nick picking, and did a bad job at it in that others corrected you. So, again. I'm not worried about what you think.

Dan

by alaniane

I think a main point missed is that programming does not have to equate to mathematics for mathematics to be useful to the programmer. Higher level mathematics gives the programmer an extra tool. You can program with just a basic knowledge of addition and subtraction just like you can build a house with just a saw and hammer. However, it sure makes it convenient to also have a level and few more tools. The same goes with programming. If you can learn the principles used in higher mathematics (whether you take them in college course or learn them on your own), it gives you an extra tool to do the job.

by deepsand

And, many engineers [u]can[/u] do theoretical work, but find the application of such to be the more challenging and rewarding.

by absolutely

The theory generally subsumes the knowledge employed in the engineering tasks guided by it.

No doubt many engineers understand theory very well, but your joke still depends on acceptance as generally true of something that is not.

by alaniane

Even the French Fry cook will have to know some mathematics if he is to survive...well maybe not!

by deepsand

calculus.

Technical analysis of such instruments uses many analogs of physical measures, such as mass, velocity, acceleration & momentum, and perforce relies on calculus for making the evaluations upon which trading decisions are made.

by tony hopkinson

What are you going on about ?

Explain the mathematical concepts that you would use to resolve this ambiguity.

Axiom of Choice, maybe?

Oy Fred, Should this be add a to b and double it or double b add a ?

Would be my pick, no program, no math !

A good developer(analyst, solution provider :D ) would go back and resolve the potential ambiguity, a good programmer would whap some parentheses in it to make the code unambiguous.

Not necessarily correct though, maybe Fred was the wrong guy to ask

or a could have been 0 :p

by alaniane

Algebra has already defined how to solve the problem. You could just memorize the operator precedence rules; however, why were those rules created in the first place?

Learning the principles behind the various mathematical rules and theorems helps one learn how to spot ambiguities in code. True, you can learn the skills elsewhere like trial and error or English grammar. However, mathematics provides an excellent basis to hone those skills.

I am not saying that mathematics == programming, it does not. However, it does not have to equal programming for it to be useful to the programmer.

by tony hopkinson

What axiom of mathematics sets out operator precedence ? That is a convention.

There is no inherently mathematical concept that will resolve the ambiguity.

You can prove where
2(a + b) <> a + 2b when a <> 0
You can prove that for any n(a + b) the result must be wholly divisible by n

If you gave me foo(a,b,c) and asked me to work out what was in it, I'd try it with a range of inputs and then map it. But I'd still need to know all three numbers were used and only once and that there was + and * in there !

You are saying mathematical reasoning will make you go back and ask which one was meant, I'm saying having a +ve IQ should do that much.

The scariest thing is a mathematician might spend a shed load of resource proving that it couldn't be proven without more information.

by alaniane

Actually, what I am saying is that math should at least teach you to think about ambiguities and solving things logically. However, math or any other subject is not going do your thinking for you. I can see your point of not absolutely needing higher math to become an accomplished programmer and I agree that you can learn analysis through other means. However, personally I have found that higher math has helped me in analyzing the specs better.
Nevertheless, math is not a cure all, you have to have some feel for what the customer wants and the ability to see where more information is needed. Math is just one tool, not the only tool that a good programmer needs.

by wayne m.

3 + 5 * 2 = 16 or 13
3 + 5 = 7
7 * 2 = 2
16 or 13 = 10

Show your work and all answers.

by alaniane

My point was that the phrase in of itself was ambiguous until a rule was laid down to clarify the equation. Most spec that I receive is a lot like that equation. You have to go back to PM or the user and ask what exactly did you mean by this. Sometimes you have to do this repeatedly and make sure you document everything. It seems to be a given axiom that the user is going to contradict what he said previously or he is going to assume that you should have known some obscure detail about his line of work that is undocumented, but generally accepted. It comes as a shock to them for example that I don't know how to drive an eighteen-wheeler, so how could I have known that this trailor size is always assumed to be the same as that size when tariffs are calculated.

by alaniane

There is some validity to the arguments above about the math needed. Basic algebra and statistics will allow a programmer to write code to solve the above problems and think most of the posters agree that these math disciplines are needed. However, I still feel that the higher maths provide a programmer with more options especially of the student can master the principles behind those maths.

by alaniane

I recently took an accounting course at a local tech college. The tests given were multiple choice and open book. This is supposed to be a college level course. Sadly, even with an open book multiple choice test, there were still some who failed the test.

by tony hopkinson

Rigour is the point of proofs, there is no rigour in software development.

If there was, we would be out of a job, or be up for a massive salary increase.

Mathematics as system of logic is based (15 ?) axioms.

There is only one axiom in software development

If you are right now, you'll be wrong soon.

by alaniane

However, it does show why some analysis should be done. Most problems that I have encountered could have been solved simply if the programmer or analyst had just thought about what they were doing. For example, I ran into one problem where calculations were off because of duplicate records in the data table. The programmer designed the table with a surrogate key to make it easier to write to it. The problem was that the surrogate key undermined the uniqueness of a compound candidate key. If he would have thought the problem through, he could have created a unique constraint for that candidate key or he could check his inserts to make sure that they were truly unique. Since he did not think the problem through, he ended up creating a new row in the table everytime someone calculated the form. Customers get really testy when there are charged over and over again for the same items.

by tony hopkinson

Didn't say the math was wrong, the model the math expresses could be though.
Pure math is an almost enclosed system, it's when you attempt to apply to the real world you get problems like

The chance of this electron being here is 90% so there must be another 10% universes in existence where it isn't.

Or because I can take line that describes my place in time and manipulate a point on it to be before any other point therefore time travel is possible.

That's what I mean by the map not being the territory, being able to express something mathematically, only means it exists as a mathematical 'reality'

by tony hopkinson

99.9999% of the time the only math involved is the size of my overdraft if I keep making daft arse 'mistakes' though. :D

A better example would have been overflows.

by tony hopkinson

to use floats in the first place?

Floats are inaccurate, you can do the math to prove it if you want, but the conclusion is all that matters to a programmer.

Single, double you are setting the precision and magnitude and accepting a level of inaccuracy.
If accuracy is mandatory, don't use them, that's what things like BCD are for.

by tony hopkinson

is a place for us with those skills, but
optimisation for footprint or speed is an increase in complexity, which is an increase in cost, which means business will draw the line on good enough.
Lets face it the first thing you should learn about optimisation is when to call it a day.

by deepsand

Do you actually believe that much code is still typed character by character? And, without benefit of an error checking editor?

TR has [b]no[/b] such tools here.

Human speech allows of a great ability for understanding despite wide variations in spelling & punctuation, pronunciation & cadence.

by apotheon

As a programmer, I know that floats are inaccurate. Being a mathematician wouldn't change that fact. As such I know that the proper way to handle accuracy with a given decimal place limit for any given calculation, with absolute accuracy necessary over the long haul, the way to handle it is to use integer arithmetic and collect modulus values until you've got enough to perform the required integer arithmetic again.

That solution requires a little algebra, a whole lot of arithmetic, and reasonable knowledge of the programming language you're using. It sure doesn't take a dozen credits or so of calculus.

by apotheon

. . . because you've run out of worthwhile arguments. Every time you disagree with me, you get stubborn as a mule and absolutely refuse to listen to anything anyone else has to say. Actually, I've noticed you do that with pretty much everyone -- but it's less obnoxious when you're disagreeing with someone with his head up his behind so far he's looking at the morning's lunch.

by apotheon

Are you aware of how unutterably self-righteous and full of yourself you sound when you start up with this crap? It's not my fault you can't summon a useful argument and have to resort to name-calling, insults, and asinine holier-than-thou proclamations.

by apotheon

This is a direct result of your self-righteous nonsense getting so absurd that I find it increasingly difficult to take you seriously.

Seriously.

by dannyaaa1

I haven't read all 609 of these arguments, but here in Alabama, you can't get an english-speaking teacher for Trigonometry or Calculus. So you have to spend another 2 years learning conversational German before you can pursue your Mathematics. Stick with the Algebra requirements!

by wayne m.

There are at least two types of analysis performed in software development, requirements analysis and program analysis. Neither is based on the type of analysis done in mathematics.

Requirements analysis is based on language, oral and written. It involves understanding of various shades of meaning, identifying and resolving apparent and real contradictions, determining rules and exceptions, and dealing with almost infinite variations. This type of analysis is better learned in a Literature course analyzing written works than in a mathematics course deriving equations and proofs.

Program analysis is even less clear cut. Here we use subjective concepts like cohesion, coupling, readability, testability, complexity, and bigness. For every problem there is almost an infinite amount of solutions and an even higher amount of dead ends. We try to eliminate the dead ends and argue endless over the nuances of "better" solutions. Where is the mathematics in determining which classes to create? Whether to use comments? To use a stored procedure or application code? I would love to have a proof to give the right answer, in practice I have to be satisfied with an answer that is accepted without physical violence or bloodshed.

In a branch of mathematics, one works in a closed world based on a set defined beliefs (axioms). Except for the extremely small few who create a new branch, mathematics involves the study and application of that branch's axioms. Software development works in the world of human people and underlying, universal truths do not exist. It is a world of many right answers, each true and each false in a different way.

I respect mathematics and I respect software development, and I understand they are largely independent.

by deepsand

In my nearly 50 years of experience in this regards, I have found that it is not uncommon for mathematics to be used for analyzing both the problem and its possible solutions, not to mention the implementation of a chosen solution.

by tony hopkinson

Their real value is they by their nature they remove a lot of extraneous detail, so you don't get someone whining about a spelling mistake, of choice of font, and missing the fact that they omitted a bit of the spec.

Nor do you suddenly find it's been sold to ten customers, even though it's not even close to being finished.

There are still a whole raft of questions that have to be answered before you can build the model though which is the point I think Wayne is trying to make.

by alaniane

Academia can give someone a background to help them to program. Besides, academia can only provide you with knowledge, you have to learn how to use it. I like one quote I heard (I forget who said it), but this is the gist of what he said, "You can live without breathing for a few minutes, live without water for a few days, live without food for a few months, but most people live their entire lives without thinking." Experience is going to teach you how to program a lot more than any academic course is.

by alaniane

I think I see where you are coming from. I concede the point that mathematical reasoning and programming logic are very different beasts. I do agree that Math != programming. However, I do feel that the one discipline can and does complement the other.

by apotheon

Formal education doesn't really teach you anything at all. It's just one way among many to make information available to you for learning. You still have to actually learn it yourself, whether you get it from a professor, a book, Wikipedia, or divine revelation.

by tony hopkinson

As much as I find that sort of thing annoying the worst code / designs I've seen have been by computer science 'professionals'.

How about a look up table keyed by varchars, read them into memory, just the keys !
Loop through a file, read in the lookup key, find it in the memory list, call a function , to query another field out of the look up table using the key value in the list.
Five fields + the key in that table, so that five queries per record in the file, plus the existence check, yes he queried the look up table to see if the key he'd queried out was still in there !

He was consistent though did the same thing on another table 20 + fields in that.

So effectively 30 ish select Field from Table where key = '' per record in a file.

What a muppet !

by daveo2000

I don't remember being one of those who receives complaints on my English. Did I miss something?

by deepsand

Is your memory that short that that's slipped your mind? Or, do you simply think that others might accept your false statement without question?

by apotheon

You're the one here who has implied insult to two people -- me, and someone not even involved in this discussion -- in the immediately previous post (previous to my comment in which I referred to your insults, that is). Good job, contradicting yourself with your [b]very next post[/b].

by deepsand

To quote you, Absolutely, "it was only a joke," one not authored by me, but one that certainly got the point across quite well. The author's assumption, an obviously correct one, was that the listener(s) would quickly and easily understand it.

by absolutely

Punch the "View All" hyperlink, and reply as [i]deeply[/i] as you wish!

"The author's assumption, an obviously correct one, was that the listener(s) would quickly and easily understand it."

I understand. I do not agree. "Therefore, what?"

by tony hopkinson

Like you I fail to see the point of writing a shed load of code that comes up with the wrong answer, and then rescue it by using my mathematical skills to calculate how inaccurate I was.

That's just stoopid.

One system we wanted the weight in metric tonnes to three decimal places. In other words kilos.

It was really difficult to write that bit of code to insert the decimal point for display and reports. :p

Course then some twit imported them into another system as a float, did a categorised report, added up his group totals and got a different answer to me.

This was communicated urgently to me as a software fault in my code, because his numbers added up!

Damn eejit.

Things are a lot better now, but you used to be able to get the same code with the same data come up with two different answers on two machines dependant on the numerical co-processor. The explanation for that would go over a manager's head at very high altitude! Quite rightly he wouldn't give a toss.

by deepsand

re. the manner in which they state their case, or the degree to which they defend against those in opposition.

So, it seems that the real problem is not the how another carries himself here, but whose "side" he's on. If he's on yours, all is well; otherwise, there's hell to pay.

by apotheon

I've corrected people several times around here even when they substantially agree with me in terms of conclusions. You just don't pay attention, I guess -- or you choose to pretend it hasn't happened. Only you can say for sure, though I'm sure you'll just go on saying it hasn't happened.

Hell, I've even agreed with the opposition from time to time. In another thread, I even agreed with rickk pretty publicly once or twice in the last couple days. Yes, really -- rickk. He said something intelligent. Who'd've thunk it?

by deepsand

My knowledge of what you do or do not say to others is based on threads/sub-threads in which we have both participated. What you say elsewhere is hardly a priority of mine.

by deepsand

And, I've better things to do with my time than switch back & forth between views just to compensate for TR's having screwed up this view.

by joules ampere

Quite often it is enough to use a formula or algorithm that can solve a real world problem without knowing how it was derived this is useful and saves a lot of time even if you have a PhD in mathematics, why reinvent the wheel? If an algorithm or formula already exist to model the problem and there is no accuracy or performance problems you should use it.
Unfortunately we are not in control of the problems we are going to encounter and from time to time a new situation crops up that you need to model yourself, this is where mathematical knowledge becomes indispensable, if you are a programmer without mathematical skills I can assure you that you are going to feel inadequate, I challenge you to find every solution in ?Numerical recipes in C? , ?Numerical recipes in Fortran?, or any other book or the Internet for that matter, it won?t happen, you will be surprised to know how many seemingly small problems can result in mathematical knowledge requirements. While this knowledge is not necessary sometimes, it can be a powerful friend to have when the time comes. I would not say you have to be a genius at tensors and calculus but a little college mathematics can go a far way in any type of programming. The universe is a mathematical entity and the problems therein. You cannot escape from using or knowing mathematics if you want to become an independent first class programmer. The objective of every teacher is to prepare the student to reach the top. If you do not grasp mathematical concepts, modeling a situation becomes more difficult and you will remain at the bottom of the programming heap using snippets of code that may not be optimal and sometimes scouring the internet in vain for that special solution, this is far from enjoyable. As an Engineer I find myself not using some branches of mathematics I learnt for quite some time now but should I had skipped learning those branches of mathematics that I knew I would not likely use? Of course not, I do not know when the requirement for them will come up, the argument posed here is not whether higher mathematics will be used or how often it will be used in programming but if a programmer should be taught mathematics, I would certainly believe so.

by tony hopkinson

:p

Proof please

This should be interesting

Given U = The universe

LMAO

Prepare the student to reach the top of what?

If it's the profession of programming they are failing horribly.

by joules ampere

You are missing the point, we have not yet discovered that one formula that governs all the forces in the universe, Einstein died trying, (See The unified field theory),or a Theory of Everything but even without these we have gathered enough knowledge to realize that mathematics is at the centre of understanding celestial mechanics, the behavior of atoms and subatomic particles, electromagnetism and so on, if you read something on astronomy or cosmology you will realize that our galaxy and others like it cannot be understood without mathematical treatment. The Plastic that makes your keyboard, the food you eat and any other material on this earth all have chemical models that can be understood by mathematics. As far as humans are concerned what we know now suggest that the universe is mathematical in so far as we need mathematics to properly describe it or at least the observable part, this should not be the point of focus, it is a little mischievous to ask for such proof, this argument was only used to illustrate how many problems in this world we find ourselves demand mathematical solutions, be it a gaming, scientific or business system, which takes us back to the core of the discussion as outlined by this analogy: There are some auto mechanics working without being taught the basic physics of the automobile, yes they can grind your valves and know enough symptoms to effect a repair on your car, they may even know how to use a dwell meter to adjust the timing and so on but if they understand a little of the physics along with their experience and rules of thumb they would be far more equipped to deal with new situations, that is what we mean by being at the top, you have a serious edge over your competitors, I am not saying the mechanic should be able to explain the Carnot cycle or the Otto cycle but if he can I would be more inclined to give that mechanic more business than the one who knows not, all else being equal. Therefore if you are running a school on automotive repair you would have as a part of your curriculum basic automotive physics it can only help. God forbid even a little mathematics may help him, maybe he will run into a velocity ratio calculation problem with a gear train. There is no good reason not to apply this same thinking to teaching programming, based on the simple fact that some problems may require an original mathematical solution. Nobody is going to tell me that a game programmer that has a good grasp of linear algebra won?t be better off than if he does not. How do you know what problems you are going to face as a programmer? You may be able to seek a solution from outside but would it not be smoother if you could solve your problem yourself? You can ride a motorcycle without a helmet for years without a problem you may even walk away from quite a few accidents without serious injury without this protection but by law in most countries you are required to wear a helmet, for a good reason, it may very well save your life, there has been no known cases where a helmet causes death. The statistics that suggest that you can become a motorcycle rider without head protection for years without harm is not used to determine whether or not you should be required to wear one, rather it is the statistics that suggest that when the cyclist is in an accident the helmet more often than not saves his life. Yes there are many who break the law and are happy riders but are they on a sound footing? I think not. The same applies to programmers that have the protection of mathematical understanding; it can be a life saver or maybe a job saver.

by tony hopkinson

The universe is an entity that has some properties that can be abstracted through mathematics.

The map is not the territory.
Other than that I fail to see how my disagreement with you one that isue, got me painted as a mathematical luddite.

How do I know the problems, I've been doing it for thirty years.
Number of times used calculus, none.
Numbers of times got messed up by, global variables, memory corruption, memory leaks, side effects, bad comments, ambiguous logic, compile depenadant logic, poor exception management, races, deadlocks, livelocks..
Too many to count.

by apotheon

The way you phrased that was an obvious attempt at insult of two people simultaneously, and anyone with half a brain reading what you posted would see the same implications I did.

by deepsand

If you don't like being compared to maxwell, then refrain from calling others "dolts," etc., or from labeling their postings as "crap." To insult others and then cry "foul" when you get flagged for such is unconscionable.

by apotheon

Your ridiculous attempts at insult were in response to me throwing your own words back at you.

by apotheon

I'd ask why you made blanket statements about my behavior then, when I corrected you, you professed disinterest in the subject, but I know it'd be futile.

by deepsand

I count 9 instances in which you've used this word as "rebuttal." Is that some sort of technical term?

by apotheon

I didn't use the word as "rebuttal". I used it as a descriptive term in the middle of a statement that, as a whole, was a rebuttal.

By analogy, I might as well say that you used the word "instances" as a "rebuttal". With argument tactics like this, it's a wonder everyone hasn't simply put you on ignore already.