views:

177

answers:

3

In the spirit of this question I wanted to have a sense of what is the proportion of time split between fixing bugs and implementing new features. If possible try to give an estimate for the product as a whole as opposed to individual developer stats and try to make an average over the course of a typical year. Do provide a general descriptive of the product/project to allow comparison. Specifically :

  • Maturity of project
  • Is it still actively developed or strictly in maintenance ?
  • Size estimate of the product/project
  • Size of team developing it (all inclusive)
  • What is your team score on the Joel test.

Ex :

  • approx 80% time spent bug fixes 20% new stuff
  • Mature software (20 years old)
  • Actively developed
  • 1.5M Line of Text, approx 700k - 900k LOC
  • 12-15 actively coding in it.
  • we got 5/12 for sure, some would say 7/12.
A: 
  • 50% debugging, 50% new code (and personally I want the debugging part to be lower)
  • Software is 15 years old
  • 1.5M line of text (with 170K empty lines, 250K lines of comment, 800K lines of actual code)
  • about 10 people developing on it
  • Joel Test: 8/12
Patrick
sounds like we are in a similar situation. What you think is the primary reason you spend so much time on bug fixing and what could you do to reduce this figure ?
Newtopian
Problems: Some parts of the software have become too complex. Software is too flexible (allows too many configurations) so we can never foresee all possible configuration combinations. Solutions: Better scope determination. Keep things simple. Test sooner. Automated testing. Make sure developers can focus on their task. Make people responsible for the quality.
Patrick
yes... this sounds familiar... The software slowly becoming a platform to do everything but end up not doing anything right without copious amount of scripting in a proprietary scripting language, that, in itself is also becoming a platform... yikes
Newtopian
+1  A: 

I would estimate: New features 70%, Bugs 10%, Technical debt (refactoring etc.) 20%

  • Maturity: 2 years
  • Actively developed
  • Team size: 8
  • LOC: 50k-100k
  • Joel Test: 9/12

You didn't ask for technology stack but if you are interested it's Ruby on Rails

AaronThomson
Thanks, yes this is exactly what I was after
Newtopian
This is almost the same as my current project. Differences are Joel test is lower, its asp.net/c#, and lower team size.
corymathews
+1  A: 

I'd estimate we spend around 70% of our time on new features and 30% on Bugs.

  • Maturity 10 years
  • Actively developed
  • Team size 14 (1 manager, 1 tester, 1 UI designer, 11 developers (8 on new features and 3 dedicated to maintenance))
  • 2.2M lines of text (950K actual code)
  • Joel Test 10/12
Big GH
Great thanks... May not be significant statistically but there seems to be a correlation between the Joel Test and the ratio. Although I am not surprised by this find it is nice to see data going this way.
Newtopian
An interesting observation regarding the Joel Test. We certainly didn't start out on this development with a 10 score (it was more like a 5), but we added the others over time which has helped with the ratio. Another thing to note is that our ratio has also changed over time. About 2 years ago we spent approx 50% of our time on Bugs before splitting the team and really trying to focus on new development.
Big GH