PDA

View Full Version : Chatter Did you write any code last month?



Phil Pilgrim (PhiPi)
12-21-2011, 02:51 AM
Or the month before? Well, if so, you'd better take another look at it. According to this article (https://threatpost.com/en_us/blogs/quality-coding-takes-break-holidays-why-122011), code written in October and November tends to be buggier than programs written in other months of the year.

-Phil

erco
12-21-2011, 02:59 AM
I resemble that remark. My buggy code is equally flawed year-round!

Unsoundcode
12-21-2011, 03:03 AM
Hmmm, are there 26 weeks in each of those months or is it just me, the question is rhetorical.

Jeff T.

doggiedoc
12-21-2011, 03:04 AM
I resemble that remark. My buggy code is equally flawed year-round!Hey!! That's what was going to say!!!

ElectricAye
12-21-2011, 03:09 AM
... code written in October and November tends to be buggier than programs written in other months of the year....

Probably a result of meeting deadlines, the end of their fiscal years, etc. But it does raise the question: could it be some sort of seasonal affective disorder, diminished sunshine? Well, it's grey in Seattle almost all year round - maybe that explains a lot about what's always happening with Windows.

Phil Pilgrim (PhiPi)
12-21-2011, 03:18 AM
... could it be some sort of seasonal affective disorder, diminished sunshine?
That's exactly where my thoughts were headed.


Well, it's grey in Seattle almost all year round - maybe that explains a lot about what's always happening with Windows.
LOL! But, you know, their software was a lot better when they were located in New Mexico!

Oh, wait. I'm a Washingtonian, too. I don't know about my software quality as the result of gloom, but I do know that my productivity declines between July and September, due to the distraction of good weather and long days.

-Phil

Martin_H
12-21-2011, 03:28 AM
I'm always skeptical of software metrics. In theory measuring code quality would be bugs measured per line of code. But organizational politics encourages gaming the metrics.

I've seen QA pad their bug counts by reporting the same bug more than once, just have a separate set of steps which converge on the same screen.

Conversely how do you define a line of code? Good code is often more compact than bad code, so a single bug has a higher ratio than multi module ramblings. Second once you measure and reward a behavior some people will change their behavior to match it.

Setting all this aside, I think planning cycles often follow the calendar year, so the schedule crunch is probably a contributing factor.

ElectricAye
12-21-2011, 03:44 AM
... I do know that my productivity declines... due to the distraction of good weather and long days...

Then vee must keep you chained in zee gloomy brick-lined vork vroom where you will be most productive! Back to vork!

http://filipspagnoli.files.wordpress.com/2008/07/john-hurt-1984-big-brother.jpg

Cluso99
12-22-2011, 08:51 PM
I always thought code announced or delivered on April 1 was the most suspect ;)

erco
12-22-2011, 09:43 PM
Actually, here's some code I wrote in November: http://www.youtube.com/watch?v=TftZYLkJcyU

It's not buggy, just quite useless. Inspired by the "you suck if you use GOTOs" crowd.

Publison
12-22-2011, 10:29 PM
I think Bruce has been chained to the vork vroom and has been most productive. :)

87904

ElectricAye
12-22-2011, 11:11 PM
I think Bruce has been chained to the vork vroom and has been most productive....

The resemblance is striking.

Tor
12-23-2011, 11:02 AM
I've never seen any such trends in the code I've written over the last +30 years.. although code I write at the end of December tend to be better, simply because I can work when nobody else is at work and thus I get more peace to stay 'in the flow' (if you're a dedicated programmer you may be familiar with what that is).

More relevant is maybe that these days I'm also involved in keeping track of all bug reports for all the software written in-house.. and I see no negative (as in "more bugs") trend for code written in October and November. I can already say that a lot of the code that was produced in-house in those two months for this year is especially good.

-Tor

stamptrol
12-23-2011, 11:54 AM
I'm with Tor on this one.

I think there is a direct relationship between buggy code and the number of interruptions per hour that the programmer is subject to.

The science shows that there is no "one minute" interruption. It can take many minutes to get back in the groove after even a simple break in concentration.

Also, I always found early morning programming (as in 7 am, not 3 am!) vastly superior to pulling all-nighters.

Cheers and Merry Christmas to all!

lanternfish
12-24-2011, 03:13 AM
I resemble that remark. My buggy code is equally flawed year-round!

Hear, hear! What would code be if it wasn't flawed?