S.O.S. Mathematics CyberBoard

One of my friends told me that mathematicns only think about algos and never think of implementing them and they never see the big picture they're more into details ? is that true?

_________________
all in all we're just some bricks in the "wall"..

I'm a mathematics student and I work as a software developer. I'd say I am a pretty good developer--I haven't had any complaints from people who use my software, so if that's any indication.... heh, maybe they just say it's poorly written behind my back At any rate, I think your friend is talking out of his/her ass. Mathematicians in general, do what is needed for the job. If an algorithm is needed to be designed, or an upper bound for the runtime needs to be found, that's what will be done. It's not necessary to implement the algorithm. I mean, a decent coder should be able to look at any algorithm (provided it is not beyond his/her level of understanding) and turn it into code. So, even if your friend's statement is true, what's the deal?

How did this come up anyway? Was he given an algorithm from a mathematician that he is having a hard time turning into code, or?

-- X Conrad X

theory and implementation does not always go together. They are not always equilvalent, and the task is usually much harder than you think. It usually can be done, but it requires skills. even translating pseudocode into actual implementations is hard sometimes.

_________________
Has anyone noticed that the below is WRONG? Otherwise this statement would be true:

where

The choice of language has a role to play in it as well (some operations are more/less awkward in different languages). Note also I said if the algorithm is not beyond his/her level of understanding. If the coder understands the algorithm and the coder has a commanding knowledge over the language, translating pseudocode into code shouldn't be a problem. I've never encountered any exceptions to this, but I don't claim they don't exist. In general, I believe what I said above holds true.

-- X Conrad X

GOD, your opinion holds well for development of "simple" systems like databases, corporate programming. If you want to translate an algorithm from a SIGGRAPH paper into code, sometimes, it's *very* far from trivial. It takes a lot of skill, thorough understanding of the algorithm, and a lot of experience to do it right.

ahhh, SIGGRAPH papers. For the summer, my work had being dealing with a collision detection and response system which was based on a couple of SIGGRAPH papers (solid/cloth/hair).

of course, the limit on the length of siggraph papers (i think they have to condense it down to 10 pages or something) doesn't help.

_________________
Has anyone noticed that the below is WRONG? Otherwise this statement would be true:

where

Again, knowledge of the algorithm and commanding knowledge over the implementation language were premises. Why is this overlooked with every reply? Christ.

-- X Conrad X

They're not overlooking it, they just don't agree with you. There's a difference.

_________________
It's quiet now; and as I think my thoughts alone,
I try to keep my head straight, but I think I'm too far gone.
For in this silence, the truth rings even louder.
A constant grinding, begging recognition of its power.

Thank you for your unwanted and yet still solicited opinion. I was referring to



He provides his "counter example" but it falls right in line with my premises! Hence it's not a counter example at all. I didn't say translating pseudo code is of equal difficulty for all scenarios, but it is relative given the experience of the programmer, his knowledge of the algorithm, and so on.

-- X Conrad X

SURE.

_________________
"The only way of discovering the limits of the possible is to venture a little way past them into the impossible." -- Clarke’s Second Law

Oh my GOD, don't get offensed so easily

I just want to make some of *your* points stronger, and provide a bit of a different point of view with some concrete examples. That's it. No counter-examples. You said something very general and I brought it to a specific example, that's all.

Your point was too general to underline the important things, because of this, I said your opinion is not right, but actually I agree with you.

It's all good.

-- X Conrad X

On a public message board? Surely you jest...

_________________
It's quiet now; and as I think my thoughts alone,
I try to keep my head straight, but I think I'm too far gone.
For in this silence, the truth rings even louder.
A constant grinding, begging recognition of its power.

Yes, all jest

-- X Conrad X

if you know exactly how the langugage works, and if you know exactly how an algorithm works, then clearly you won't have any problems implementing an algorithm in that language except for typos when implenting! so clearly making those assumptions is unreasonable. Hell, even a robot could do it, if it knew exactly what to do (written out in pseudocode step by step)!!

In actuality, you must provide your own implementations in a paper, which is based on numerous other results! And clearly a lot of details will be left out and it is up to the implementer to decide. And clearly, initially nobody except the arthor has "commanding knowledge" of the algorithm, so it is all part of the process to learn the algorithm, decide on the best implementation, do it, debug, and finally produce.

GOD, when you have time, go pick up a set of Donald Knuth's The Art of Computer Programming.

_________________
Has anyone noticed that the below is WRONG? Otherwise this statement would be true:

where