Fog Creek Software
Discussion Board




Review of Joel's lecture at UCDavis

I know Joel's blog pointed to Ted Neward's review at TSS.NET.  But a much more technical review is available over at Ted's blog at:

http://www.neward.net/ted/weblog/index.jsp?date=20040129#1075429554816

Dilip
Friday, January 30, 2004

"As part of that last point, he asked the audience, "If a bug costs $100, do you fix it if it will cost $110 to fix it?" Most everybody said no. "What if it costs $90 to fix it?" More hands went up. Then he pointed out, "The problem is I haven't given you enough information--what if that same $90 can go towards advertising, and brings in $300? From the business perspective, that's obviously the better ROI for the money." And a few engineers in the audience laughed, until they realized he wasn't making a joke."

This excerpt from TSS made me recall the discussion we had here (sorry, can't seem to find the topic in the archive) on development and business priorities in the light of the time spent fixing the huge upload problem of CityDesk.
Has Joel changed his mind?

Just me (Sir to you)
Friday, January 30, 2004

Out of curiosity, at some point in Joel's "cost of bugs" discussion, did he point out that buzz on one $100 bug can kill your product?

Philo

Philo
Friday, January 30, 2004

I would say that if the "buzz" on that bug "kills the product", then that bug cost a lot more than $100.

Foolish Jordan
Friday, January 30, 2004

How is the cost of bugs being determined? By programmer hour?  If a programmer costs $100/hr, a $100 bug takes one hour to fix? Does a $90 bug takes 55 minutes to fix? And a $110 bug takes 65 minutes to fix? Is there any real difference between 55 minutes and 65 minutes?

Setting specific prices on bugs makes it appear that Joel is "telling like it is. " He's setting hard values on bugs. But the problem is, if you think about it more, his numbers have no meaning.


Friday, January 30, 2004

>> "This excerpt from TSS made me recall the discussion we had here (sorry, can't seem to find the topic in the archive) on development and business priorities in the light of the time spent fixing the huge upload problem of CityDesk."

I believe the article was on software craftsmanship.  I don't think adding the ability to upload large files while allowing the user to do other things is fixing a bug.  It is merely adding a convience feature to the product.

Dave B.
Friday, January 30, 2004

I think the point of the "meaningless numbers" is to make people realize that outside of academia, there is more that goes into deciding to fix a bug, than the fact that bugs are bad so you should get rid of them.  If fixing the bug delays your product 2 months so your competitor gets to the market first and capitalizes on 2 months of being the sole market player, than maybe fixing the problem where 1 in 100,000 data inputs crashes your program.  I don't know if the MS approach of release and patch is the best solution, but it gets them the money, and outside of academia that's what people write software.  For the income.  No one was ever a full time software developer for free that wasn't already wealthy.

Elephant
Friday, January 30, 2004

But software must be free! People who code for money are just monkey coders. They do not love what they do. All good developers like their work so much they are willing to do it for free to help the gerater community. We must stop the selfish acts of those who seek to destroy the profession by charging money. if we all instead work to 'help each other' out, that is the only true path.

Zealot
Friday, January 30, 2004

Elephant, point taken. However, you are setting up a straw man by mentioning academia. Nobody cares about fixing bugs in academia.


Friday, January 30, 2004

That is so true. It is absurd that Joel is basing fixing bugs on economic decisions. That explains why commercial software is of such poor quality compared to free software. in the free software community, someone will fix the bug without charging for it. Therefore the dollar cost of fixing all bugs is zero and therefore it is always economically advantageous to do the fix. In the greedy microsoft world they use economic arguments and hence their software is very buggy, unstable, and riddled with security holes. Also, in the free software world there is no money spent on marketing. This also opens up time and resouces for coding bug fixes. If anything, Joel's speech only proves that free software is the only way!

Richard Stahlmain
Friday, January 30, 2004

Richard, the dollar cost of fixing bugs in an open source environment is not zero.  This is basic economics.  Time spent fixing a bug is time that could be spent doing something else.  There's the risk that a bug fix will cause other bugs elsewhere.  There's the risk that a bug fix will add extra complexity to the code.

A relevant and well-known quote from Jamie Zawinski: "Linux is only free if your time has no value."

Junkster
Friday, January 30, 2004

>>Richard, the dollar cost of fixing bugs in an open source environment is not zero.  This is basic economics

Basic economics would say bug fixing costs money if developers charge for it. If the fixing is done in the developer's spare time, or as an exercise by a college student, it's hard to define this as costing money to anyone, especially if the bug in question is not causing your business any harm.

Logically enough, we should substract bug fixing to the GNP. Do we? Considerning how crude economics is, I don't understand why we even bother trying to evaluate the cost of software development...

FredF
Friday, January 30, 2004

Please do not be naive.

T.J.
Friday, January 30, 2004

1) this 'richard' guy is probably a parody of RMS.
2) many people do not understand 'opportunity cost'.
3) opportunity cost is not always a simple dollar calculation.

mb
Friday, January 30, 2004

Nobody's time is ever free. We all have expenses -- housing, food, heat, water, and so on. At a basic level, you could sum those up, divide by the average number of full time work hours in a year, and that's hourly cost for someone to live. To avoid going bankrupt, that person would have to work for an hourly rate of at least that much.

If you accept that a person must have enough money to live before they're able to fix bugs in Linux, then you must accept that Linux isn't free. In many cases, people have simply donated their time instead of money to fund it's development.

Joel's right. There's a cost to doing work. Every feature has a price. Every bug fix has a price. One way or another someone pays.

Jeremy
Friday, January 30, 2004

Joel makes an obvious point when he assigns specific dollar numbers to bugs and ROI, but that ain't reality. The reality is that you don't know, exactly, or even roughly. It comes down to wisdom and experience when making the call on which bugs to fix...but that doesn't make for good lecture talk.


Friday, January 30, 2004

Every decent manager makes this estimate before scheduling a bug to be fixed. Whether or not the estimate is accurate is another story.

pdq
Friday, January 30, 2004

"Linux is only free if your time has no value."

Often, your time has no *MONETARY* value.  One of the most common traps people fall into is the idea that monetary value is the only value that exists or matters.  The old saying "time is money" is a perfect example of that.  In reality, time is ONLY money when you are "on the clock" (or paying someone who is).

Time has other values besides monetary value.

if you are paying me $20 an hour, then an hour of my time is worth $20.  However, if I spend an hour working on some software, and I'm not getting paid, the first question you have to ask is "what would I have been doing if I wasn't working on that software?".

If the answer is "watching TV" then the time I spent on that software has no monetary value -- it didn't cost me anything.

If I spend a few hours a week patching Mozilla bugs  instead of watching "Everybody Loves Raymond" then those bug fixes are essentially free -- zero monetary cost.

It doens't mean that my time has no value -- it just means that my time has no *MONETARY* value.  But many people have trouble grasping this concept because they are too locked into the idea of "time is money".

Officer Barrone
Saturday, January 31, 2004

So time spent coding instead of with my family, does that not cost me something?  Time spent coding instead of getting outside and going hiking, does that cost me nothing?  Time spent coding not going out and being sociable with friends, does that cost nothing? 

Time may not cost money, but don't kid yourself, it's not free, as there is only so much of it to go around.

Elephant
Saturday, January 31, 2004

Great points, OB.

The economic value of time expended for something must be considered relative to what the time would have been spent on otherwise.

When a person voluntarily chooses to work on an open source project, that in itself is good evidence that the time spent doing that work is preferred to whatever else might have been available to do at the time.  As a result, the time so spent may even be considered a positive economic gain for the person whose tastes lead them to choose to develop free software.

T. Norman
Saturday, January 31, 2004

Elephant, if you always value your time with friends and family more than you value time spent developing open source, then you won't develop open source.

For someone who *voluntarily chooses* to develop open source software, their values are different -- they may prefer to spend 10 hours a week programming open source and 35 with friends and family, instead of 45 hours with friends and family.  For them, the 10 hours a week isn't a cost because the prefer it over 10 hours doing something else.  Otherwise they wouldn't do it.

T. Norman
Saturday, January 31, 2004


Um...The guy pretending to be Dick Stallman is just trolling..And you all fell for it.

Go back to the interesting discussion of bug vs cost stuff.

Duh
Saturday, January 31, 2004

Quick! Somebody grab the net!

Richard Stahlmain
Saturday, January 31, 2004

*  Recent Topics

*  Fog Creek Home