Of Rockstars and Bricklayers

Today we're going to start with a thought experiment.  You are one of the better coders in your company.  All the proof I need is that you're reading this.  Ok, that may be a little self-serving but it is true that the better coders I know are always reading something about coding.  It might be a new language, or a new technique, a programming question and answer site, or even a site that likes to rag on bad code. 

Based on how many Code Monkeys you know, it's easy to figure yourself in the top half of programmers, a 'good' coder if you will. An assumption that is easy to make is that the higher your productivity the higher your pay.  Is this true?  If you base it on simple experience, the answer is no.  It's too easy to get a job right out of college and get mediocre raises and then find yourself interveriewing candidates for openings in your own group where you have to offer 'market rates' to attract anybody.  If you've been there for 5+ years, given a company desperate to hire warm bodies, it doesn't take much for the new guy to be making 15-25% more than you even if he is a two-finger typist and can't spell SQL if you spotted him the 'S' and the 'Q'. 


Now, I'm happy where I am but it didn't take much to trigger my bad habit of thinking up 'what-ifs'.  The Stackoverflow podcast has been a good one for this sort of pondering.  Jeff and Joel have enough real-world experience to bring a ton of wisdom to the discussion on programming as it's own topic and in episode #42 discuss underperforming programmers.  I'd been reading about how star actors can command huge salaries based on their past work and the expected effect on the box-office if their name appeared on the marquee.  That major-league sports can have individuals who make 10-times what the lowest paid athlete makes on a single team, it doesn't take much to ask why programming doesn't see the same effect. 

I submitted this as a question to the Stackoverflow podcast which they spent a good 10 minutes on in episode 77. It didn't fit so well as a question on Stackoverflow itself, I suspect the culprit is the lack of code snippets. Interestingly enough I discovered the podcast when they were at episode 65 and decided to start listening back at #1 and was up to around #42 when I decided to contribute my question.  So anyway, when I saw the summary for #77, I didn't have the audio handy so jumped over to the transcript wiki for #77 and read the transcript.  It was a pleasant surprise at how engaged they were and that the question wasn't dismissed out-of-hand.  When I downloaded the audio, to hear their discussion it only reinforced my reaction.

The range of talent can be mighty wide even if there is some disagreement on the empirical numbers.  Joel Spolsky talks about the difficulty in finding great programmers, and uses similiar metaphores in a comparison with opera singers. Jeff has a similar take on Coding Horror.  Finding great talent is hard because there are only so many people who can do what you need done.

There have been some good discussions that has surfaced.  John Cook has a good argument based on lack of meaningful metrics.  It's a question that has been around for awhile.  Frank Wiles makes the claim that hiring average (or below) programmers actually costs your company money (followup post). Obviously when a company says they only hire the to 10%, they are deluding themselves or at least paying lip-service to wanting good programmers.  Joel does a great job of shattering that particular illusion. 

What about sales? In some fields top salesmen can bring home seven-figure incomes so we know that the wide range of incomes aren't limited to Hollywood and the big leagues.  James Kwak picks up on John's post and supports Joel and Jeff contention that paying the top performers much more than the average, outside of easily measured areas like sales and movies, has a corrosive effect on the group and the best coders end up leaving for greener pastures while the dregs hang on for dear life or are eventually pushed out.  I can't disagree.  It seems that if you are underpaid you have to ask yourself Ann Lander's question, "Am I better off leaving or staying put?"  Just realize that it's no use complaining to the company, the status quo works for them.

Life's Too Short To Work For Clowns.

One of the cliche's of our culture is the little boy who rebels against his parents and dreams of running away to join the circus. His mind is filled with the glamor of the Big Top, the exotic animals, the independence of wandering the country from one town to the next. He hasn't really thought about what it means to work in a circus. The Ringmaster did many menial jobs before arriving in the spotlight. The animals have to be cared for; double the time it takes to feed them when you consider both ends of the animal. Traveling from town to town may sound romantic until you realize it means you spend more time traveling than in one place, lots of lonely nights on the road regardless of the weather or climate because 'the show must go on" after all. The life of a circus is not just popcorn, parades, and pretty girls.

I ran across a quote in Christopher Duncan's, "The Career Programmer: Guerilla Tactics for an Imperfect World" that made me both laugh and sigh.

You need to consider one additional thing when trying to convince management the current approach to developing software needs improvement. What do you do if non of this sways them and they insist on continuing with chaotic development approaches that consistently create disasters and stress? Update your resume. Life's too short to work for clowns.

So the question becomes, do you work for clowns? Does the development process at your company make a circus look like a well-oiled machine? You might work for a clown.

I plan on putting together a complement to The Joel Test where instead of a list of things a company does before you consider working for them, The Clown Test would be a list of things that indicate you might work for clowns.

This might describe your boss,

• if he thinks an estimate is a promise,
• if he thinks developers are interchangeable with no effect on the schedule,
• if he thinks its ok for the lead developer to publicly berate a junior developer,
• if he thinks requirements documents are a waste of time,
• if he thinks the only place for designers is at a fashion show,
• if he thinks he just needs one good programmer who can code anything,
• if he thinks that people whine too much about teamwork,
• if he believes in Silver Bullets,

he might be a Tech Clown.

My apologies to Jeff Foxworthy.

When you run across a good example of a Tech Clown, send 'em my way and I'll add it to the list for every one's enjoyment.

The Myth of the Aging Developer

I question that I've heard throughout my career is, "What happens to older software developers?" The danger with this question is when the discussion turns to the implied version which is "The lack of older developers means that it is a young persons game." Nothing could be further from the truth..

My wife and I are both developers. We have both been working as developers for 15+ years. One thing we noticed was that there were a lot of people working as developers who really shouldn't be, they didn't have the aptitude, interest or sometimes even the background. It would be easy to dismiss the problem as a lack of proper education but that doesn't explain the developers who lack a CS degree but who can code circles around veteran devs. Those folks usually decide that programming is not for them and move on to different job. Some go into management, some become analysts, while others change careers completely.

What's left? The ones who are passionate about development and are interested in doing it right. They love learning new stuff and don't gripe about how things used to be in "The good ole' days." They might be grumpy but they usually know their stuff.

The main problem would be in finding management who doesn't have a pre-conceived notion about what makes a good developer. If they feel that only the younger ones have the stamina to keep up with the pace, I'd wonder whether those same employers were the same ones who believe that the 'Deathmarch' style of project management is standard operating procedure. They'd rather grind through fresh-meat developers (Code-Monkeys anyone?) than listen to the experienced ones and change their development process much less change their management style.

Lastly, I've noticed that doing nothing but coding holds less business value than being able to communicate with customers, users, management, and vendors. One reason you start coding less is because finding answers to ambiguous requirements takes more experience and provides more payback to the business.

My advice, to those who worry about their coding future, would be to find a place where they need experienced developers to help with the overall task of creating or maintaining software, that also gives you the opportunity to code regularly. That way you can have the time to keep up with industry changes while taking advantage of your understanding of software development.

What I learned about Cryptography

In preparing a presentation on cryptography for a graduate class, I was surprised by what I learned.

Cryptography has been used from ancient times through the Industrial Age to the modern day. It's been used for military, diplomatic, criminal, and commercial purposes.

In the simplest and oldest form, Greeks used strips of paper wrapped around sticks of a certain size called a scytale - rhymes with Italy.

Julius Caesar is one of the first known users of the simple letter substitution method. The modern day cousin is ROT13.

If he had anything confidential to say, he wrote it in cipher, that is, by so changing the order of the letters of the alphabet, that not a word could be made out. If anyone wishes to decipher these, and get at their meaning, he must substitute the fourth letter of the alphabet, namely D, for A, and so with the others.
—Suetonius, Life of Julius Caesar

Hiding messages (steganography) is as old as the ancients. An example would be a tattoo on a shaved head. The hair would grow back hiding the tattoo. A modern example is hiding text in the pixels of a digital image.

Charles Babbage is a name taught in every into to computer science class as a pioneer in computing machines. I never knew that he helped the English during the Crimean War by breaking cryptological codes.

Mary "Queen of Scots" was caught when one of her secret codes were broken.

The Voynich Manuscript may have been an elaborate hoax to defraud a king, who had a interest in mysteries like coded messages, out of a large amount of gold.

A mob boss used encrypted messages to give out orders to his lieutenants and was arrested when the police cracked his amateur code.

Cryptography is thousands of years old and is used all around us everyday.

Programmers - Poets or Mathematicians?

Programming software is much like writing a story. Any given story can be told in a near-infinite number of ways and those who practice the craft will spend just as much time, if not more, arguing over the style of a work, as if their subjective opinion can suddenly become objective fact. Usually this leads to a contest of "If you don't agree with me, I'll just keep talking until you see reason" or otherwise known as talking the subject to death.

I thought I'd shine some light on an example of this issue in programming. There is a never-ending debate in programming circles about the 'proper' way to write code that creates some true/false decision. Stackoverflow has a good example of this.

My personal pet peeve (petty but my teeth grind everytime I see it) is verbosely setting booleans, e.g.

bool isValid;
if (percentage >= 0 && percentage <= 100)
  isValid = true;
else
  isValid = false;

whats wrong with

bool isValid = percentage >= 0 && percentage <= 100;

It's soooooo much more succinct and easier on the eye

Even though the single line version isn't the most complex, different people read through code in different ways. A single expression can be easily parsed by people who have a math background or are experienced coders. That's not the point. The expanded version reads more like a sentence than a formula. It's easy to forget that there are plenty of coders without a background in math. It really comes back to the preferred style of the reader. Do they prefer reading expressions and formulae or something more like human speech.

Homer's The Illiad and The Oddessy are ancient works dedicated to expressions of speech. They were originally poems that were passed down in an oral tradition and meant to be sung instead of read.

Euclid's Elements is a work dedicated to expressions of logic. It covers objective proofs and unambiguous facts and is the beginning of the mathematical tradition, if you will, of rigorous scientific thinking.

The problem we face in software is code is a form of human expression, with all the imprecision that entails, attempting to communicate with the utterly logical CPU; programmers will be forever trying to translate "An Ode to the Machine" into the Pythagorean Theorem.

If Gulliver Were a Designer

The term 'Design' gets thrown around a lot in software developer circles. There are constant flamewars over favorite techniques and battles over what good design really entails. When those arguments start getting too ad hominem figure out whether the speaker is a Code Monkey before giving them much credence. Since I've already braved the waters, see "Have Design Will Travel" or "The Age of the Design Review", instead of safely avoiding the issue, I'm going to bravely (or foolishly) join the fray.

Realize that there is more than one type of 'Design'. Design in the large turns into architecture and is something that really only comes with experience. Designing at the lower levels, say at the individual class/object level is easier to cover. Today I'm writing about design in the small vs the large; Lilliput vs. Brobdingnag.

The issues with designing a class is the same regardless of platform or language. The key is whether an object should be autonomous or whether it is better for any given behavior to be spread out among objects with limited scope and distributed responsibility.

For each class the answer might be different. We end up with a spectrum along which we can place classes based upon the Density of Responsibility.

             (Level of responsibility for behavior)
        Autonomy - - - - - - - - - - -  - - Dependence
     High
 C      -   [god object]                [spaghetti]
 l      -
 a      -
 s      -                                
 s      -           
        -                  
 s      -
 i      -
 z      -
 e      -  [template]                   [framework]
      low

Let's say you favor letting the class perform all the behaviors itself, or as many as you can. Starting on the left side of this graph, when you make your class more autonomous, the size of the class will grow unless you continuously refactor it to make it more generic. This leads to a template. If no refactoring is done, the tendency is for the class to become more "god-like" because if there is some behavior it needs, it has a method for that - it can do anything and everything. The number of fields and methods grow and soon become both unmanageable and unavoidable. Since the class already does so much, coders would rather add to the monstrosity than try to piece it apart and cut the Gordian knot.

The right side of the graph has classes that depend on other classes to a large degree. If the dependency level is high but the individual class is small, that is a sign of a framework; each class doesn't do much and requires lots of dependent classes to accomplish some function. On the other hand, a highly-dependent class that also has a large amount of code is a sign that the class is full of spaghetti.

The key to this question is to determine where you feel more comfortable on the graph. In any event, individual classes will end up spread out on the graph unless some organizational principle is applied, which is how you can achieve the results of Template or Framework.

Having just written that, I would say that there is a correlation between class size and degree of organization. Robert C. Martin (or "Uncle Bob") covers similar ground with package dependencies in his very thorough paper on Design Principles and Design Patterns[pdf]. JDepend is an implementation of the ideas behind the graph on page 26 and complements static analysis tools such as Checkstyle and PMD.