views:

358

answers:

8

I'd like to re-ask the How to justify a workstation upgrade question that Garry Shutler asked with the context of some of the excellent answers given.

I believe that quantifying the productivity increase (extra working hours) is the most effective way to do this.

My case in point: I have a fast machine at home and a slow one at work. My estimate is that I would gain about 30 minutes a day of extra productivity at work if I had my home machine at work. This productivity would come from less waiting to do all the tasks that I do. (An extra 30 minutes a day is about 3 weeks a year.)

Problem: I need to measure this.

Is there a software utility that can monitor and scientifically quantify time taken by tasks on a machine?

+1  A: 

Excel and a stopwatch?

chills42
+2  A: 

Something from Sysinternals might do the job they have coded a nice fancy task manager which may be able to help.

Obviously your work machine may have twice as many tasks running and communicate with a DB over a WAN and your home PC doesn't. Both would drastically affect the perceived speed of a PC. From then it just gets more complicated but being able to accurately monitor every task would be a great start.

Here's a link to the process monitor, you might just like it anyway :)

http://technet.microsoft.com/en-us/sysinternals/bb896645.aspx

Robert
+7  A: 

Break it down into pieces you can quantify. For example, I compile every 4 minutes and each compile would save 10 seconds. But after a while I get bored of waiting to seconds so I go to Stack Overflow and I'm there for two minutes. Sometimes I'll start talking to Jim and that takes 4 minutes for both of us. So 15 times an hour * 8 hours * 10 seconds = 12 minutes + 5 trips to stack overflow = 22 minutes + 4 conversations with Jim = 38 minutes for me and 16 for Jim.

The next step is to see whether it's worth it to buy a new computer. Let's round it off to an hour a day and you cost the company $100k per year in salary and benefits. One eighth of your hours are wasted so that's $12,500 per year in wasted productivity between you and Jim that could be saved by getting you a faster computer.

But you're not going to throw the computer out. The boss's new admin doesn't need a brand new PC, and buying her a computer would cost $1,000. Your computer costs $3,000 so it's really only costing the company $2,000.

It's not difficult to make it look like a no-brainer. The goal is to put it into dollars, but of course that doesn't guarantee you anything.

Dan Goldstein
It is such a sad thing to do, but unfortunately is the only way to do it, otherwise the mgmt does not realise.
+1  A: 

There is one simple reason why to get a faster machine:

  • Reducing frustration
  • Working more efficiently

Every time something hangs, you sit there waiting - being distracted from what you where doing/thinking.

Waiting time is dead time. Even if the system just hangs for a second, you still notice and get distracted..

Edit: I don't see compile time really as a problem. Any decent CPU can compile your stuff in under 10 seconds (on a not toooo complicated project). I see more of a problem when you wait for stuff to come up or shut down.. Starting stuff should be fast..

Tigraine
That's two simple reasons :)
Guy
Wow, I'd like a project that compiled in 10 seconds. I think you may be surprised if you surveyed the size of code bases that many developers are working on.
Wedge
@Guy: Our two weapons are fear and surprise...and ruthless efficiency!
Pistos
<10 seconds is a good aim for incremental builds, after changing a small amount of code. Usually a function of how the build is set up, not just computer speed.
postfuturist
10 seconds build in C++ is a dream...
Constantin
A: 

On *nix, I'd use time when doing builds and such.

Snagging the output of that is very simple, and easy to compare machine to machine.

Also - you can time certain "non-work" things, for example a build of GCC or KDE or something else fairly big from source.

Additionally - there have been some related questions (can't find them atm) on multiple monitors: make sure you add that in, too :)

warren
+7  A: 

I am sorry, but I believe if a manager or CEO needs actual numbers to justify a high-end machine for high productivity worker, then you have bigger problems at hand!

Ather
Amen brother ;o)
wcm
+2  A: 

You might look at the book How to measure Anything for some tips on how to get estimates. While you're asking for a more recent machine, you might also ask for dual monitors. There's hard research that shows it improves productivity. You should also consider that your costs are a lot higher than your salary, maybe twice as much.

It's always odd when you have a faster machine for your personal use than at work but it happens. Finally, if you can't get any satisfaction and it really bothers you, look for another job. If you can't get another job, figure out what you have to learn to be more desirable to other employers.

David Nehme
Thanks for the book recommendation - looks good - I've just ordered it.
Guy
A: 

Make sure you don't spend more time justifying your needs, than a new computer would gain you.

jsl4980