Does anyone know ut开发者_开发问答ility which can measure work intensity. For example - keystrokes\mouse clicks per second, time spent in IDE\other developer tools, total lines of code per hour...etc.
It's widely held that any measurement of such type is not accurate. More often than not, it can be more productive to delete code than to add it. It can be more productive to think and plan than type.
I suggest you measure work intensity via the meeting of goals. If goals are met, and work is good (tested, etc) then it follows that the work was done.
(source: osnews.com)
Not only for code quality!! Can be used for work performance too..
Check out Timesnapper. I haven't gotten around to using it yet, but it looks really interesting. Basically it takes a screenshot of your desktop every now and then, and can produce graphs of how much time you spend in each application, among many other things.
Simplest way to monitor lines of code is with a source code versioning system. Subversion has triggers that lets you log the commits, lines, etc. and even notify the team.
As to the other metrics, I think they are BAD for measuring developer work performance. Results are what matter, and that comes out in the final code checked in, and the application quality.
I employ several developers and one guy will write 50 lines of code in short bursts of productivity, and the other guy plods along. What matters is results.
Plus, staring at the screen while trying to figure out a problem is productive, but nothing is moving.
I have made a program like that years ago to measure when I work too much and for personal use only. My opinion is that if you are search for something to measure your work so you can ask for money, or you are the boss and search for a way to measure your workers, this is not the correct way, and I have try that on me...
There is a program that I know and use, the todolist, with the ability, to messure the time that you spend on each task. How ever I have years to count the time spend, now I check the results, how fast programs work and run, how robust they are, and not product bugs, if its easy update, if they get me back money and other stuff like that, and not measure the clicks and what ever...
http://www.codeproject.com/KB/applications/todolist2.aspx
I have meet a coder in the past, that make on macromedia director a project, and make a solution with 10.000 frames, its was impossible to update it, buggy ugly etc, and then one other make the same project with 100 frames, and lingo code. The first one work for 2 months, the second one work a week to design it and a month to code it and product a far better project...
The first one did not actually know to programm, and the second one never explain in the first one that didn't know what he did because there are competitors.
So how can you measure a case like that ? by the way, the boss of this 2 programmers, think that the first one programmer works most, and actually work most, day and night because the project the way they fix it needs to copy paste again and again many thinks. The company close after some years, for other reasons, and this project that I mention, was a demo that never actually released. I was a friend of one other person inside this company that just go for a coffee there and just show me this project - I make my comments to my friend.
精彩评论