What is in your opinion the time span you can effectlvely work?

[Thaddy]

As a social scientist - one of my hats - , I am curious about what, in your opinion, is the time span you can effectlvely work per day?
I mean, it is my experience as a - retired - development manager, that the average time-span differs a lot, but generally is at most 5 hours after which productivity drops off steeply unless a lot of adrenaline is involved for just a few times a year a dead-line needs to be met or a bug-fix needs to be released a.s.a.p. In any other case, being present does not make sense: better to send them home and start afresh the next day.
Do you have the same experience? Not about other people, but about you or your team?
What is already clear from other research is that programmers, just like medical docters and astronauts, have a very high brain-strain while at work, so I want to find out if my assertion makes any sense.

In my own case it differed between 4-5 hours median and over 24 under very occational stress.
The assertion is that presence at work in hours does not make sense and it not beneficial to an organisation: the organization would loose money because the median is skewed and does not represent average.

[dbannon]

Depends on the supply of Pizza and Coffee !

Seriously, my team were often in the office around 8:30 to 9:00 am, finished 4:30 - 6:30 pm and would be logged on as soon as they home quite a lot. But they loved what they were doing.

Thats the difference, enjoy it, its fun. Don't enjoy it, its a drag.  My super power ?  Get people doing the job they enjoyed.

Davo

[Thaddy]

Davo, that is the issue: keeping time does not make sense. Getting the work done makes sense and is not related to average presence but to median performance. That is an important difference I want to investigate: the sillyness of presense.

[Weiss]

if we are talking specifically about programming, I would say, 2 hours on average. But that really depends on the type of programming I do. There are tasks which are highly creative, where I get involved entirely, it is like being asleep almost. Then I can go for 5-6 hours in such highly focused state. Boring tasks on the other hand, like cleaning up my code and tying up lose ends, this is where I take a lot of breaks and work in bursts of solid 15-20 minutes of work, followed by a quick two-hour breaks in-between. I am hobbyist, this question was always on my mind, how do people do this all day.

[Weiss]

probably need to add, as a new to programming, especially at the beginning, I would roll in bed at night figuring out how code works.  I would be writing code before brushing my teeth as soon as  get up. Things like that. I sometimes would feel the "pump" in my head. Same as muscle pump (if you lift, you would know), but in the head, a sweet tiredness, mixed with endorphins from job well done. Not so much now, but at the beginning it was like that.

[Thaddy]

I would roll in bed at night figuring out how code works.

Yes. I provided for that in my model, because that is often implied. Same as smoking pause can be very productive. Better do it in your sleep and the next day solve it straight away.

[Zvoni]

Speaking from my own experience (Excel-VBA, No Programming Team, Solo Jockey):
I agree with Thaddy that it doesn't make sense to keep Time.

I have days, i get maybe 20 lines of code done, because the problem escapes me completely.
Other days, i get 5 Classes, and 3 Standard-Modules done.

BUT: Independent of the day described above: I always make a short coffee or smoke-break on average every hour for about 5-10 minutes, NOT thinking about the coding-issue.
It's to relax my mind

[Thaddy]

My wife and my guitars do that, not smoking.

[Zvoni]

My wife and my guitars do that, not smoking.

Well, i don't have a wife nor children (at least, none that i know of....), and i don't drink....
so.... it's smoking for me to get me "down" again... 

[cdbc]

Hi
Well, all of the above, actually... Add to that, I roll my own cigarettes, in batches of 7, that little break, lets me do a tedious task, while the brain wanders and more often than not, comes up with an idea or a solution.
Keeping time is useless. Time doesn't matter when you're in the 'zone' and it matters just as little if you're not!
Consider this: I'm Danish, my girlfriend is German(we speak german at home), we live in Denmark and I program solely in English...  One could say, my brains gets enough exercise 
Regards Benny

[MathMan]

Thaddy,

I did professional programming long time ago - nowadays I'm only a hobby programmer.

Totally concur with the statements from the others here that single times highly depend on the circumstances, but as you asked about average I can provide the following.

I'm on a large task which will occupy me for years to come and I'm on it since several years now. I faced all situations that have been described by others - abandoning the whole thing for several months, even a full year in one case; very complex stuff to implement & debug (large part of asm routines), rewriting things from scratch, etc. When counting everything up, including comments, I'm on a run rate of ~60 new lines of verified code each day. Dividing that by the purported magical figure of 10 LOC per h, I would be on average 6h a day - remarkably close to your 4-5 h a day.

Cheers,
MathMan

[Martin_fr]

I think my record is somewhere at about 42 hours on a day. I let you know when I break the 48 hour barrier.

[Zvoni]

I think my record is somewhere at about 42 hours on a day. I let you know when I break the 48 hour barrier.

You unraveled the secret to Time-Compression/Dilation at non-relativistic speeds?

[MathMan]

I think my record is somewhere at about 42 hours on a day. I let you know when I break the 48 hour barrier.

You unraveled the secret to Time-Compression/Dilation at non-relativistic speeds?

Who said he wasn't moving at relatvistic speed 

[440bx]

When I was much younger, I got into the "zone" after about 10 hours of programming.  I routinely programmed for 24 to 36 hours, mostly non-stop (gotta occasionally eat and shower.)

One time I wanted to make a point about how much time it would take to complete a particular project.  That time I programmed 76 hours in about 80 continuous hours (I had to keep track of the time because it was billable ) 

I couldn't do that today even if my life depended on it.