[skipjack]

Top 10 lies of Engineers


1. ¿We're about to go into beta testing.¿ This is a meaningless statement because it doesn't matter when you go into beta testing--what matters is when you come out of beta testing. (The only hard and fast deadline for coming out of modern-day beta testing is ¿before you run out of money.¿)

In the good old days, ¿alpha¿ used to mean ¿all features are
implemented though not necessarily working properly.¿ ¿Beta¿ used to
mean ¿there are no more repeatable bugs.¿ Nowadays beta means ¿we've
gone as long as possible past the shipping date that we promised our
investors.¿

2. ¿I don't know anything thing about marketing...¿
This is a lie of false modesty. The engineer is thinking, in totality,
¿I don't know a thing about marketing, but how hard could it be
compared to what I'm doing? I should run marketing and engineering. I
just hope that the marketing the MBAs come up with is worthy of my
code.¿ However, don't worry too much about this lie because it
self-corrects as the engineer misses deadline after deadline and comes
to realize that he has bigger issues.

3. ¿I'll comment the code, so that the next person can understand what I did.¿
This is a lie of good intentions. Really, the engineer did intend to
comment the code but as the schedule slipped, priorities changed. The
question put to management became: ¿Do you want me to comment the code
or finish it sooner?¿ Guess what the answer was. Luckily, the lack of
comments usually doesn't matter because the code is so crappy that a
total rewrite is necessary in a year.

4. ¿Our architecture is scalable.¿ This is the lie
that I enjoy hearing the most. Typically, an engineer who has never
shipped a product says this after creating a prototype in Visual BASIC.
The whole conversation goes like this: ¿Google's architecture isn't as
scalable as mine. They can support 25 million simultaneous searches. We
will be able to easily handle a billion.¿

Luckily, in most cases, the adoption of the product is slower than
the CEO's ¿conservative¿ forecast, so scalability never becomes an
issue. Yeah, those clowns at Google, Yahoo, Oracle, Microsoft, Apple,
and AOL don't know anything about scaling compared to the engineer...

5. ¿The code supports all the industry standards.¿ This is almost a truth but for a short omission: ¿This code supports all the industry standards that I agree with.¿
The engineer has made a personal decision to ignore standards she
doesn't like--for example, those promulgated by Microsoft. It's no big
deal--customers will never know...

6. ¿We can do a Macintosh version right after we finish the
Windows version; in fact, much of the Windows code can be re-used
because of how we architected it.¿ The truth is that version
1.0 of any software is an experiment. It can be a magnificent
experiment, but it's an experiment nonetheless. Thus, Windows version
1.0 is held together by duct-tape. The Macintosh version is a copy of
the duct-taped Windows version written by an engineer who just finished
college and got his first Macintosh a month ago. How hard could it be
to learn to program for a different platform? C++ is C++, right?

7. ¿We have an effective bug reporting database and system.¿
Of course, the assumption behind the design of the bug reporting
database and system is that there are no bugs in the code, so there's
not much to database and report. Generally speaking, if the largest
number of documented bugs doesn't ever exceed 1,000, it means that the
company isn't tracking bugs carefully.

8. ¿We can do this faster, cheaper, and better with an offshore programming team in India.¿
Rank and file engineers usually don't tell this lie; it's the CTO who
does. Somehow we've got it in our heads that every programmer in India
is good, fast, and cheap, and every programmer in the United States is
lousy, slow, and expensive. My theory is that for version 1.0 of a
product, the maximum allowable distance between the engineers and
marketers is thirty feet.

9. ¿Our beta sites loved the software.¿
In twenty five years of working in technology, I've never heard a
company report that its beta sites didn't like its software. There are
three reasons for this: first, many beta sites are so honored to get
pre-release software that they don't want say anything negative.
Second, most beta sites haven't used the software very much. Third,
most beta sites don't want to seem cruel by criticizing a company's new
product. Doing so is as socially unacceptable as telling someone that
his baby is ugly.

10. ¿This time we got it right.¿
The scary thing about this lie is that the engineer really believes it.
Again. The problem is that ¿this time¿ occurs over and over again. I
have great faith in engineers and believe that in the long run, they do
get it right. It's just that in the long run, we're all dead.

Addendumbs (sic):

¿This code is so bad it would be faster to write it all from scratch than debug and expand the current shipping code.¿ (Joel.)

"I like thinking about architecture, but I can code." (Glenn Kelman)

"It works on my machine." (Gaurav)

"Of course I can let go of the code and run the business instead." (Jason)

"Even my mom can navigate the screens." (Nitin)