[ishkabible]

so, after coming back to reality for the holiday, i realized i had neglected to do a weeks work for my intro to computer science class and that it was all due by 12:00 tonight(i finished at 11:45!!). now in my cramming and thoughtless cranking out of work i found a question. it asked "the average programer can write, test, and document ____ lines of code per day" it then gave the options, 10, 15, 20 and 25.

i was shocked. i thought there must be an issue. i guessed 25 as it seemed closest to my guess of 100. turns out the correct answer was 20! my guess was based off the fact that i can write about 300 LOC of working code in 1 day in C++. given another 2 days i can work out most of the bugs in this code, and at most 1 more day to write/generate some documentation and add any other comments. that puts me at 75 per day and a more experienced developer could probably do better, so my guess was 100. how can i be so far off of reality? i have no experience in a real work place as a programer so i figure i must be missing something here; something big apparently. what is it?

This post has been edited by ishkabible: 27 November 2011 - 11:01 PM

[tlhIn`toq]

Where is this question coming from, and where is its answer coming from? Because in my world 20 lines a day is B.S. That would get me fired after a week.

Let me know which companies find that acceptable so I can put in a resume. I'd be lead developer in no time with that sort of pitiful excuse for competition.

[ishkabible]

well, the only source i have to site is my class's book "New Perspectives on Computer Concepts 2010". they site some study i would hope

my book is at school right now; ill find what study/companies they site if any tomorrow

This post has been edited by ishkabible: 27 November 2011 - 11:15 PM

[Lemur]

Lines that work a day? Properly documented/commented? If I had the 8-5 block I could easily go over 1000... I've done that in C# in about 4-5 hours easily. Meh, maybe I just type fast.

[Nikitin]

Lemur, on 28 November 2011 - 01:09 AM, said:

Lines that work a day? Properly documented/commented? If I had the 8-5 block I could easily go over 1000... I've done that in C# in about 4-5 hours easily. Meh, maybe I just type fast.

Lemur, you are my most favorite member on this forum.

[Lemur]

I do seek to entertain. To be fair, anyone types fast in a programming test worth 35% of your grade with a limited timeframe. Do multi-line comments count in that number?

[no2pencil]

A count of Lines Of Code is no measurement of effectiveness or achievement.

if(condition) continue();

Is comparably the same as

int i=0;
i=if(condition);
if(i==value) 
{
 continue();
}

[Lemur]

Depends on how much condensation is done, if it's still (for the most part) readable or at least commented it's fine. Heck, most of the shell scripts I see are people trying to fit as much as they can into a single command.

[AdaHacker]

ishkabible, on 28 November 2011 - 01:57 AM, said:

how can i be so far off of reality? i have no experience in a real work place as a programer so i figure i must be missing something here; something big apparently. what is it?

The big thing you're missing is how this number is measured. It's not like somebody checks your code in the morning, then comes back at the end of the day to see how much you've written. It's usually an aggregate value for a project, not a measure of how much new code a particular developer can write in 8 hours.

Without more information, it's safe to assume that this number is calculated at the end of the project - the number of lines of code in the final system that were written by the developer divided by the number of working days the project ran, or something similar. So this includes time spent debugging, refactoring, etc. Depending on how it's calculated, it may even include time not spent coding at all - e.g. system design, documentation, etc.

It may also help to consider that the kinds of projects that keep these metrics tend to be large, complicated, and process-heavy. This can make a big difference in productivity. A developer might spend an entire day tracking down a bug only to fix it with a 1-line change. Or he might spend the day refactoring a section of code, reducing the number of lines and thus giving him negative productivity for the day. All that stuff is going to figure in to the average.

[lordofduct]

Lemur, on 28 November 2011 - 08:01 AM, said:

Depends on how much condensation is done, if it's still (for the most part) readable or at least commented it's fine. Heck, most of the shell scripts I see are people trying to fit as much as they can into a single command.

Man I hate it when water droplets form inside my code... makes the semi-colons bleed and run and end up turning into OR pipes.

That's why I keep my code at a stable 78 degrees Fahrenheit and in a dry room, no condensation for me.

[jon.kiparsky]

AdaHacker, on 28 November 2011 - 09:51 AM, said:

ishkabible, on 28 November 2011 - 01:57 AM, said:

how can i be so far off of reality? i have no experience in a real work place as a programer so i figure i must be missing something here; something big apparently. what is it?

The big thing you're missing is how this number is measured. It's not like somebody checks your code in the morning, then comes back at the end of the day to see how much you've written. It's usually an aggregate value for a project, not a measure of how much new code a particular developer can write in 8 hours.

Without more information, it's safe to assume that this number is calculated at the end of the project - the number of lines of code in the final system that were written by the developer divided by the number of working days the project ran, or something similar. So this includes time spent debugging, refactoring, etc. Depending on how it's calculated, it may even include time not spent coding at all - e.g. system design, documentation, etc.

It may also help to consider that the kinds of projects that keep these metrics tend to be large, complicated, and process-heavy. This can make a big difference in productivity. A developer might spend an entire day tracking down a bug only to fix it with a 1-line change. Or he might spend the day refactoring a section of code, reducing the number of lines and thus giving him negative productivity for the day. All that stuff is going to figure in to the average.

This.

Add in meetings (I've seen days eaten alive by meetings), absences, the back-and-fill when you realize that you have a line of development that's not going to work, learning the domain, finding and learning the required libraries, and those 3000 LOC days average out to a pretty low number.
But that's the real world. Schools a little different.

[ishkabible]

on page 675 the book this came out of has 1 sentence that says "Research has shown that, on average, one person can write, test, and document only 20 lines of code per day." i can find a source that is cited; the only thing i can find is where they cite there pictures and small blurbs as coming from it must be measured out of long projects, from conception to release.

this book has given me more than a few headaches. it seems to focus more on random facts and what it calls "lingo" or "the talk of the trade" rather than "concepts" that it claims to impart. i guess im glad im getting this stupid coerce out of the way now so i don't have to take it later.

[lordofduct]

Sounds to me like yet another useless peace of statistical trivia in the grand scheme of things.

I guess a manager could maybe use this number as an over-all arc statistic dealing with planning. But really, they would need to know WHY that statistic is so low and what it actually means. So in the context of your book, and your focus of study (which probably isn't management), it's useless.

What's the text book? Maybe it's a book geared towards managers and leads. Where "lingo" and "talk of the trade" is the big thing because those jobs are more about perception and looking good. This doesn't make them any less of a job, engineers need a smiling face to project their image out.

[ishkabible]

"New Perspectives on Computer Concepts 2010" im taking it from a community college as it is the only way i can take the class while still in high school. i think it's more geared more toward the technical school crowd honestly; only 2 people, of which i cam one of, in my class even plan to get more than associates degree.

This post has been edited by ishkabible: 28 November 2011 - 09:33 AM

[lordofduct]

oops, just noticed you posted the title near the beginning of the thread.

Yeah, looks to me like more of a lesson for the normals. More like a tool useful for assisting non-computer users in slightly understanding heavy computer engineers. Like a German phrase book for an American visiting Hamburg. "DID YOU KNOW: More sausage is eaten in Hamburg than actual hamburgers... har har har."

This post has been edited by lordofduct: 28 November 2011 - 09:42 AM