...Why scientific programming does not compute

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 686138502
RAC: 568690

RE: But the point is back

Quote:

But the point is back to the beginning in the OP and the original follow-ups -- scientists are rotten programmers who do not know just how rotten they are because they have taught themselves. Sort of like a self-taught hostage negotiator when you think about it.

This is just a polemic generalization that, like most generalizations, does a big deal of injustice to a whole community of people.

So how many lines of scientific code have you seen, personally?

Anyway there are many scientists out there who team up with professional software engineers to check or improve their codes. Science, when done well, is a collaborative endeavor that involves many disciplines.

CU
HB

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284713501
RAC: 106859

Well I've been a

Well I've been a non-programmer professional, but who has 'programmed' and gone through a number of phases. First as a hobbyist just before and during the IBM clone breakout, as a 'regular business user' for brand name products ( OS's and apps ), and more recently [ after 3.5 years of a tertiary course - I did all the undergrad units at RMIT in Melbourne, but haven't bothered to claim the degree ] a 'neophyte programmer' who knows enough to be quite dangerous.

So with the provisos implied by that background ( which I divulge so that others can gain proper perspective from my comments ) I would say/advise :

- be professionally taught. Get a good institution ( RMIT is widely regarded as being in DownUnda's top three, and was so placed by the industry's 'professional collegiate' ie. the peak professional/industry body said they were good ). Pay them what they want. It is worth it. This lifts you from the hack/build/debug cycle, and the mindset that typically goes with that : you are 'clever' because you have 'mastered' that cycle. Effort does not equate to progress, and one can frequently be one's best enemy ( ie. pride ). Little did I realise that like hooning/burn-outs/circle-work* in a car park, that ain't proper driving. Less marines-like hoo-ra ( or boo-ra ) and more intellect.

- programming isn't equal to coding. Coding is ( the later part of the ) translation of some 'castle in the air' into a pretty specific implementable form. The harder part is the air castle, so it can actually help to stay away from the keyboard for a while ( high level design tools aside ) and think in a language independent way - although you of course need to have a sense of what can be done, with varying 'tool emphasis', within the languages on hand.

- have professional supervision to some degree of your output. There is always someone who knows more than one's-self, so if you're lucky enough to co-opt then listen to any/all feedback.

- consider/take up the best 'self management'/'error avoidant' strategies and work practices of professionals. You don't have to do all of 'The Mythical Man Month' but one should at least read it and branch from there. Programming will result in error, but make them little ones, and even more important identifiable and localisable.

- I ( think I ) see what Bikeman means about collaborative. The 'scientist' knows the problem space better ( air castle ), the 'programmer' knows the translation/implementation side better. Talk together.

Cheers, Mike.

( edit ) Actually one very useful 'professional perusal of output' in a limited sense is compiler messages. Set the warning level to the pickiest and satisfy them all, or in lieu have a well considered reason for the breach. These compilers are written by some extraordinarily smart bods and the messages are from them to you, about your code. So it's not just a remark about suitability for consumption by computer circuitry. Of especial note is feedback related to typing issues ( not key presses, but choice of data model ) - mismatches, conversions etc - as I have found them to point out trouble brewing as regards conceptual/design problems of significance ( not merely syntax difficulties ).

( edit ) * Strictly analogous only. I of course have never done any of that .... ;-)

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Matt Giwer
Matt Giwer
Joined: 12 Dec 05
Posts: 144
Credit: 6891649
RAC: 0

RE: RE: Who cares the

Quote:
Quote:

Who cares the price of oil? If you talked to an oil exec certainly he told you that the problem was burning off the natural gas for no way to licquify and ship it off for sale.

The cost is the cost. That is the way the world works. You cannot screw with the cost by legislation.

The oil company exec's point was that with the current cost of oil it is not cost prohibitive to go thru the capturing, compressing, storing, shipping, etc ,etc process of capturing the flared oil. As oil prices go up that will become more cost effective and something they will do. They do have the technology currently it is just not cost effective to use it. As you say 'the cost is the cost' and if it cost 10 bucks per gallon, it is not something they will do right now. I have no idea what the actual costs are, we were just talking and I tried to not get into the specifics of the process or its costs, that is his area of expertise not mine.

All true but it has to go up enough so that creating/expanding the entire infrastructure for using the gas can be paid for and still make a profit.

Considering liquid only (Russia is shipping the gas at pressure to Western Europe) that is

* Capturing at the well
* Liquefying at some point
* getting it to a sea port
* building the ships and containers to carry it
* building facilities on the buyers' side to off load it
* building distribution at the buyers' side

It would not surprise me to learn it is all proceeding at the fastest possible rate and that it has been for a decade or more and that it will take a couple more decades before it can all be recovered.

Add one more bullet, pray there is never an accident that causes a pressure container to up in flames in port else the entire enterprise will come to a halt for the bad press.

Matt Giwer
Matt Giwer
Joined: 12 Dec 05
Posts: 144
Credit: 6891649
RAC: 0

RE: RE: But the point

Quote:
Quote:

But the point is back to the beginning in the OP and the original follow-ups -- scientists are rotten programmers who do not know just how rotten they are because they have taught themselves. Sort of like a self-taught hostage negotiator when you think about it.

This is just a polemic generalization that, like most generalizations, does a big deal of injustice to a whole community of people.

So how many lines of scientific code have you seen, personally?

Anyway there are many scientists out there who team up with professional software engineers to check or improve their codes. Science, when done well, is a collaborative endeavor that involves many disciplines.

CU
HB

I have seen more than I would like to remember having seen and saw coffin sized packing crates of punch cards for a single program back in the good old days. I have also written a fair amount of what could be considered scientific code back in my self-taught FORTRAN II days -- before it became FORTRAN 68. In the last ten years most of my programming has been the scene description language of Povray. However back in the 1980s I was responsible for the software development for the AN/SQS53C which was described as the largest software project in the Navy short of its space program in the early 80s. You can prefix a BS in Physics to all the above.

I believe I am reasonably familiar with the subject matter of the article. But as I was getting into that project with all the software my "advisors" had to teach me what IBM and others had sponsored during the 70s in converting programming from an art to a science. After they educated me I could lead them.;)

Where the article said in use for 30 years that is when I learned it. I had sort of assumed everyone else did. The article says the opposite.

I thought the article appropriate to all the @home projects and posted it to those in which I participate. Not meant to start a pissing contest. At Milkyway the project leader replied saying that project is a collaboration of astronomers and computer scientists and believed the issue was being adequately addressed.

More to the point if the head of the project felt it worth a response I think the legitimacy of the post has been established.

Matt Giwer
Matt Giwer
Joined: 12 Dec 05
Posts: 144
Credit: 6891649
RAC: 0

RE: ... ( edit ) Actually

Quote:

...
( edit ) Actually one very useful 'professional perusal of output' in a limited sense is compiler messages. Set the warning level to the pickiest and satisfy them all, or in lieu have a well considered reason for the breach. These compilers are written by some extraordinarily smart bods and the messages are from them to you, about your code. So it's not just a remark about suitability for consumption by computer circuitry. Of especial note is feedback related to typing issues ( not key presses, but choice of data model ) - mismatches, conversions etc - as I have found them to point out trouble brewing as regards conceptual/design problems of significance ( not merely syntax difficulties ).

As a linux user I instantly adopted yum to avoid the dependency hell of compiling source code but still do compile a few from the CVS repository rather than use the yum update. I find even in the most mature and apparently flawless apps the most common non-fatal compiler error is type mismatch. C is incredibly prone to that which is why I stick with Pascal in its Turbo incarnation even today rather than learn enough C to stop being dangerous. Single, double, quad mismatches are very dangerous things to have in mission critical apps.

Allowing a byte to be interpreted by the calling code is a recipe for disaster unless it is FORTH where you are supposed to know about such problems but then I have only heard the claim back in the early 80s that a major commercial app was written in FORTH but non-disclosure prohibited yadda, yadda, yadda.

The point is on these compiles of C source is that they are mature apps. They function as expected. They have been around for years. And still no one has bothered to correct type mismatches. Or they crop up faster than they can be eliminated. Or they only work because of the mismatches -- and that is not what one wants in mission critical software -- it only works because it should not work.

Cover your ass. It is about top get bit.

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828153831
RAC: 203591

RE: So how many lines of

Quote:
So how many lines of scientific code have you seen, personally?
CU
HB

I have a friend that used to work for a bank, a LARGE WorldWide BANK, and she used to be one of their on-call problem solvers. She would complain ALL THE TIME about programmers that did things 'differently' but still got the job down. She would have to go on line and scroll thru hundreds of thousands of lines of code looking for why the ATM, for instance, stopped working right. She would find it and then block out huge sections and fix it with a few dozen lines of code and then be made that some 'idiot programmer' couldn't code properly! Apparently there are MANY ways to skin a cat and many ways to code a program, none of which is right or wrong, they are just longer then they need to be or too short to cover all eventualities! But hey the program works, RIGHT?! Computers are getting faster and that means they can scream thru the code even faster so making 'good coding' is not important any more, just do it and make it work! My friend is OUTSTANDING at what she does and has little patience for other programmers who don't do it the 'right way'.

Me personally I couldn't program myself out of a wet paper bag and I used to program in Basic and C++ but have long forgotten everything due to my moving on to other things. I used to hate it when those darned little paper rolls used to tear as the machine was reading them!!!!

tullio
tullio
Joined: 22 Jan 05
Posts: 2118
Credit: 61407735
RAC: 0

I remember taking a bunch of

I remember taking a bunch of punched cards from Trieste University to the only computer in town, I think a UNIVAC. I was accompanying a blond girl and we took an icecream on the way down. The pack fell down and we reassembled it as we could. Maybe a Nobel prize was lost because of an icecream.
Tullio

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828153831
RAC: 203591

RE: I remember taking a

Quote:
I remember taking a bunch of punched cards from Trieste University to the only computer in town, I think a UNIVAC. I was accompanying a blond girl and we took an ice cream on the way down. The pack fell down and we reassembled it as we could. Maybe a Nobel prize was lost because of an ice cream.
Tullio

Did you marry her? ;-))

tullio
tullio
Joined: 22 Jan 05
Posts: 2118
Credit: 61407735
RAC: 0

RE: RE: I remember taking

Quote:
Quote:
I remember taking a bunch of punched cards from Trieste University to the only computer in town, I think a UNIVAC. I was accompanying a blond girl and we took an ice cream on the way down. The pack fell down and we reassembled it as we could. Maybe a Nobel prize was lost because of an ice cream.
Tullio

Did you marry her? ;-))


Unfortunately not. She was an intelligent girl and married someone who had more money. Jane Fonda said "There is nothing more erotic than money".
Tullio

mikey
mikey
Joined: 22 Jan 05
Posts: 11889
Credit: 1828153831
RAC: 203591

RE: RE: RE: I remember

Quote:
Quote:
Quote:
I remember taking a bunch of punched cards from Trieste University to the only computer in town, I think a UNIVAC. I was accompanying a blond girl and we took an ice cream on the way down. The pack fell down and we reassembled it as we could. Maybe a Nobel prize was lost because of an ice cream.
Tullio

Did you marry her? ;-))


Unfortunately not. She was an intelligent girl and married someone who had more money. Jane Fonda said "There is nothing more erotic than money".
Tullio

There is probably some Scientific formula to predict that but that is not good! You lost out on BOTH a Nobel Prize AND didn't the girl either!!! Oh well, at least you have walked on a glacier!

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.