Debugging #1 Challenge Faced by Programmers

[Brett_Tabke]

[eetimes.com...]

Sixty-three percent of developers and IT decision makers consider debugging to be the most significant problem that they encounter, almost twice as many as any other task, according to the results of a survey conducted by system simulation technology provider Virtutech Inc.

[JerryOdom]

Glad to hear I'm not the only one. Determinism/bug reproducibility tends to be a major problem for me because of the sheer number of possible inputs I usually have to deal with.

[lexipixel]

Brett, how does this apply to Perl?

Did I miss something? The article / survey appears to be of electrical engineers and developers of embedded systems (DSP's and the like) and multi-core systems.

In context of WebmasterWorld and perl/web development, I'd say the #1 "perl for web" problem is UA and UI compatability issues for the myriad of devices that now offer web access (but these usually boil down to issues of reworking HTML code produced by Perl scripts).

My vote for #2 would be security issues (which many times are OS and UA dependent and require work-arounds).

[surfin2u]

Not putting the bugs in in the first place that should be the priority, as opposed to finding them after the fact.

Not sure why this is news and why it's in this forum.

[herb]

As Dilbert is heard saying when his boss tells him he needs a SQL DataBASE to track bugs in their new CRM product. "What color would you like it". To which the boss replies, "I think green, it has more Ram and we can use it when we do the version II upgrade".

Not sure why this is in this forum

 
 

[Mopar93]

What "bugs" me the most are the programmers who are so eager to release their work "knowing" there are bugs in the program.

When I write a program, I don't even release it to a beta tester until I've fixed all of the "known" bugs.

Fix it first guys.

-Maurice

[wackybrit]

This is why Test Driven Development is the only way to go with anything that actually matters :)

[aleksl]

surfin2u: Not putting the bugs in in the first place that should be the priority

said non-programmer.
don't remember who said this:

The six billion people of the world can be divided into two groups:

1. People who know why every good software company ships products with known bugs.
2. People who don't.

and here's classic:
Every time you fix a bug, you risk introducing another one.

[LifeinAsia]

surfin2u: Not putting the bugs in in the first place that should be the priority

Tsk, tsk! If programmers didn't include bugs in the first place, they'd be out of a job as soon as the program was completed. A few well placed bugs = job security. Especially when the "fixes" create new bugs that also need to be eradicated.

[dmorison]

"Schroders Bugs" are the worst. The ones where the introduction of debug code stops the bug. I hate them.

[rocknbil]

I've found preventing bugs is more of an understanding of user psychology. "Don't introduce bugs," well, can you predict all the possible ways an end user can break your program? Just when you think you've got it, someone who thinks outside of your "box" throws some combination of actions at you, actions they can't begin to describe so you can repeat it, and poof there goes a sale.

I have spent more hours trying to figure out how someone is getting to the bug-point than actually fixing it, it is maddening.

[surfin2u]

One of the best ways that I know of for not creating bugs in the first place is to be patient and methodical, and resist the temptation to take shortcuts. For example, testing all code changes on a development platform prior to moving the change into the production system is a good rule to follow. Sometimes a seemingly trivial change is made and the temptation to skip the development test before going live arises. Learning not to give in to that temptation will prevent many bugs from happening in the first place.

There is also the temptation to write code before the problem has been sufficiently analyzed and the solution designed and considered. Coding is fun, but that other stuff is tedious for some folks. Resisting the "rush to code" is another way to avoid creating bugs.

[gregbo]

One of the best ways that I know of for not creating bugs in the first place is to be patient and methodical, and resist the temptation to take shortcuts.

In an ideal world, this is great advice. For most projects, pressures to get to market quickly force code to ship with bugs. Sometimes, the best one can do is hope that there are no show-stoppers in the release.