Is Continuous Delivery agile?

Keif Morris thoughtfully compares/contrasts Agile to Continuous Delivery (CD).

Of course the answer to the question posed in the post title is Yes, Continuous Delivery is agile. Very agile, in fact.

Keif raises several concerns traditional Agile has with Continuous Delivery. I admit to being from the old school, XP, and am a little nervous around CD. But I embrace its spirit.

With Agile leaving small release messes to clean up each iteration, or for Waterfall one giant mess to clean up in the release phase, is bad for the nerves too. At least CD picks up after itself in an ongoing basis.

As a daily Java programmer Maven makes a hash of this as Keif points out. But as with many things related to Maven, you gain a few pains and lose several others for net betterment. How does CD treat the snapshot ailment?

ESR on "C" portability across time

ESR posts on portability of "C" across time. Most times you port across machines and operating systems. "C" is rare in making straight-forward the task of porting from the past into the future.

sed is Turing complete

My command line tool is cooler than your command line tool. Thank you, Peteris Krumins, for linking. And don't forget his book.

UPDATE: A testament to my typing skills, the original title of this post was "sed is Turning complete".

Mythryl: Another great language introduction

Another great language introduction, this one for Mythryl:

Howdy! I’m Cynbe, lead Mythryl developer.

Why do I do it?

Let me tell you.

I spent the first four years of this millennium doing eighty-hour weeks at a Fortune 5 company in a division internally famous for producing more revenue per employee than the IRS.

I remember arriving home at three AM Christmas Day, sleeping thirty-six hours straight, and then driving right back to work.

It was a cool trip in its way, but over time the stress does get to you. By four years in, vomiting blood in the wee hours was starting to seem entirely normal.

It was time for a change.

By then I had written well over a million lines in C plus substantial amounts in other languages. I felt ready to take it to the next level.

“A language that doesn’t affect the
way you think about programming
is not worth knowing.”
—Alan Perlis

So I looked around to see what was new and improved. I’d learned APL, assembly, C, Fortran, Lisp, Pascal, Smalltalk, Snobol, SQL and so Forth in the 1970s, but after that there had been a long dry spell. C++, J, Java, Perl, Python, Ruby, sure, but they hardly catapult us into a new era of butterflies and rainbows. They did not expand my mind like Lisp and Smalltalk.

Happily, mostly-functional programming languages had just reached the Ready For Prime Time point.

My favorite was SML/NJ, from the nice folks who gave us the laser, the transistor, and Unix.

Unfortunately, it was research-grade code cloaked in academic jargon which hadn’t seen an end-user release that millennium.

Fortunately, I was looking for something to do.

So I set about hammering this magnificent raw material into a modern production quality open source software development platform.

To my mind Mythryl deftly combines C speed, Lisp power, and Ruby convenience with the critical new ingredients of Hindley-Milner typing, state of the art generics and just the right level of side effects.

I’m in love!

Real Options

Shane Hastie interviews Chris Matts and Olav Maassen on Real Options, an agile technique to improve IT decision making.

Do read it all. A taste:

InfoQ: Please can you briefly explain Real Options. 

Olav: Options have value, options expire, never commit early unless you know why.

Options are a way of looking at decision making. Financial options work in very narrow areas, in the financial marketplace they entail paying a fee to defer making a choice about purchasing a share or financial instrument until a later date. Our thinking about Real Options was inspired by financial options, but they are much broader in application.

Real Options are more to do with the psychology of how people make decisions than with the choices available to them. 

InfoQ: Please explain. 

Chris: People hate uncertainty, so much so that they would rather be wrong than uncertain. In a “rational preference” the hierarchy of decision making would be:

1. Have the right answer
2. Be uncertain about the answer
3. Have the wrong answer

The reality is we prefer to be wrong rather than be uncertain, so the hierarchy is:

1. Have the right answer
2. Have the wrong answer
3. Be uncertain about the answer

We can tell this because when people are faced with uncertainty, they would rather make any decision, even if it is the wrong one.

Generally people would rather have a definite wrong answer than be unsure about an answer.

Real Options is about understanding when to make a decision, rather than how to make decisions. By adding the “when” to the decision making process we remove the uncertainty and enable people to make better decisions.

Like financial options have a fixed contractual expiry date, real options have a conditional expiry date.

Jline moves forward

One of my favorite Java libraries, jline, released 2.6 recently. Among the goodies is support for ~/.inputrc. This is as close to readline as Java gets.

The science of cloud computing

The physics arXiv blog at MIT Technology Review posts The Hidden Risk of a Meltdown in the Cloud.

As with anytime you bet the farm on a technology, try to have a backup plan.

re: Romans, rubies and the D

Kajetan Rzepecki posts Romans, rubies and the D, a clever comparison of Ruby missing method look up and the compile-time equivalent in D. (It's short, please read.)

This is a clean example in favor of proper macros (not the "CPP" sort).

I've long been impressed by D. It makes up for many of the faults in C++, which impresses me less as time goes on. What is impressive about D? I appreciate that the first thing it says about itself is:

D is a systems programming language. Its focus is on combining the power and high performance of C and C++ with the programmer productivity of modern languages like Ruby and Python. Special attention is given to the needs of quality assurance, documentation, management, portability and reliability.

And that, therefore, this comes only second:

The D language is statically typed and compiles directly to machine code. It’s multiparadigm, supporting many programming styles: imperative, object oriented, and metaprogramming. It’s a member of the C syntax family, and its appearance is very similar to that of C++.

That says something good about taste in my book.

Showing test failure name for JUnit parameterized tests

A popular complaint against JUnit is lack of human-readable names for parameterized test cases. There are several solutions, this is mine.

/**
 * {@code ParameterizedTest} simplifies parametered JUnit tests by including a "case label" to
 * identify the failing branch.  (JUnit only provides an index).  The failure message format
 * becomes "descriptive string from JUnit: case label: assertion failure message".
 * <p/>
 * Fully preserves the stack trace of failing tests.
 * <p/>
 * Example: <pre>
 * public class MyTest {
 *     &#64;Parameters
 *     public static Collection<Object[]> parameters() {
 *         return ...; // See JUnit documentation for parameterized tests
 *     }
 *
 *     public MyTest(final String caseLabel, final X x, final Y y) {
 *         super(caseLabel);
 *         this.x = x;
 *         this.y = y;
 *     }
 *
 *     &#64;Test
 *     public void shouldWork() {
 *         // Body of test method ...
 *     }
 * }
 * </pre>
 * When {@code shouldWork()} fails the failure message is more meaningful, including the
 * {@link #caseLabel "case label"} in the JUnit error message.
 *
 * @author <a href="mailto:binkley@alumni.rice.edu">B. K. Oxley (binkley)</a>
 */
@RunWith(Parameterized.class)
public abstract class ParameterizedTest {
    private final String caseLabel;

    /**
     * The JUnit test rule (test extension) which fails bad parameterized cases but provides a
     * {@link #caseLabel case label} identifing the failed test branch.
     */
    @Rule
    public final TestRule parameterizedTestRule = new TestWatcher() {
        @Override
        protected void failed(final Throwable e, final Description description) {
            final AssertionError x = new AssertionError(
                    description.getDisplayName() + ": " + caseLabel + ": " + e.getMessage());
            x.setStackTrace(e.getStackTrace());
            throw x;
        }
    };

    /**
     * Constructs a new {@code ParameterizedTest} for the given <var>caseLabel</var>.
     *
     * @param caseLabel the case label, never missing
     */
    protected ParameterizedTest(@Nonnull final String caseLabel) {
        this.caseLabel = caseLabel;
    }
}

Testing thread affinity

Peter Lawrey posts on thread affinity with Java. This is a post full of numbers, good numbers. If you write low latency code in Java, this is for you.

More git: moving modules between repos

Not the coolest merge, EVER! but much more grokable, Greg Bayer provides Moving Files from one Git Repository to Another, Preserving History.

The problem: I want to move repo-A/module1 to repo-B/module1 while preserving history. Bayer to the rescue. Thanks, Greg!

How to introduce a new programming language

The Julia folks provide show the way to introduce a new programming language:

We are greedy: we want more.

We want a language that’s open source, with a liberal license. We want the speed of C with the dynamism of Ruby. We want a language that’s homoiconic, with true macros like Lisp, but with obvious, familiar mathematical notation like Matlab. We want something as usable for general programming as Python, as easy for statistics as R, as natural for string processing as Perl, as powerful for linear algebra as Matlab, as good at gluing programs together as the shell. Something that is dirt simple to learn, yet keeps the most serious hackers happy. We want it interactive and we want it compiled.

(Did we mention it should be as fast as C?)

Thanks to Michael J. Simoni for pointing out Julia. More here.

Losing malloc, gaining the world

At work we've dropped the malloc that comes with GCC in favor of tcmalloc from Google. Our creaking, monolithic C++ backend gains 10-20% performance, unsurprising given the history of malloc.

Concatenative Programming

Jon Purdy on concatenative programming, which I had never heard of before. Basically it is purer functional programming for those of you who think Haskell is too corporate. (Actually, I enjoyed the read.)

P.S. — A remark from a clever fellow, no slouch he: That made my head hurt and reminded me of, when I was a freshman, trying to do a few chapters of Abelson & Sussman on my HP 28S calculator, which had RPN and (I realized) first-class functions.. I miss RPN calculators.

Beating the PyPy drum

Laurie Tratt beats the PyPy drum in Fast Enough VMs in Fast Enough Time. Want to know more about VMs? Writing your own language or mini-language? Read on.

Here comes Kotlin

Here comes Kotlin and Guice! Thank you Alex Tkachman.

Type Theory for the Uninitiated

Kalani Thielen presents The Algebra of Data, and the Calculus of Mutation, or Type Theory for the Uninitiated. Fine calculus a bit soft, algebra a little dull, programming a walk in the park? Type theory is for you.

1-1 beats N-M

SpiderMonkey, the JavaScript engine in Firefox, is moving towards a 1-1 threading model, away from a N-M model: JSRuntime is now officially single-threaded. Let me explain by looking back.

In Solaris the threading model is M-N, that is, for each program with N user threads there are M kernel threads servicing the program. This was thought to be the most flexible design and kernel threads were a limited resource.

Linux eventually settled on a 1-1 model, that is, for each program with N user threads there are N kernel threads each mapped uniquely to a user thread. This was found to be the simplest to code and the most performant in practice.

Back to JavaScript:

A single SpiderMonkey runtime (that is, instance of JSRuntime) — and all the objects, strings and contexts associated with it — may only be accessed by a single thread at any given time. However, a SpiderMonkey embedding may create multiple runtimes in the same process (each of which may be accessed by a different thread).

I read this to say: 1-1 mapping from SpiderMonkey to user thread. The same simplifications and performance gains Linux saw from the 1-1 model will be gained for SpiderMonkey.

Old lessons learned again; better than the alternative.

Spawn of Java

Brian McCallister writes Java Daemonization with posix_spawn(2), or as I like to think of it, Spawn of Java. The problem description:

The traditional way of daemonizing a process involves forking the process and daemonizing it from the current running state. This doesn’t work so well in Java because the JVM relies on several additional worker threads, and fork only keeps the thread calling fork. So, basically, you need to start a new JVM from scratch to create a child process.

The traditional way of launching a new program is to fork and exec the program you wish to start. Sadly, this also fails on Java because the calls to fork and exec are seperate, non-atomic (in the platonic sense, not the JMM sense) operations. There is no guarantee that the exec will be reached, or that the memory state of the JVM will even be sound when the exec is reached, because you could be in the middle of a garbage collection and pointers could be all over. In practice, this would happen exceptionally rarely, at least. Charles Nutter has talked about this problem in JRuby as well.

Visit Brian's post for the solution.

(You may also enjoy his related POSIX from Java post.)

Major code irritants

I could not agree more with this Java anti-pattern: Storing money in floating point variables.

In my interviews of Java candidates (or C++, for that matter) I always include questions about creating a hypothetical Money class, and pay close attention to their ideas for representation. International candidates often fare better on not assuming US dollars, but most all candidates fail completely on avoiding floating point.

Defaulting message fields in protobuf

Not well documented is the technique for defaulting message fields in protobuf. Say you have a field of message type, named "start" here:

message Complex {
    required int64 real = 1;
    optional int64 imaginary = 2 [default = 0];
}

message Vector {
    required Complex start = 1;
    required int64 length = 2;
}

How to you provide a default value, say at the origin? Like this:

import "google/protobuf/descriptor.proto";

extend google.protobuf.FieldOptions {
    // Pick the field number that is right for you!
    optional Complex complex = 50000;
}

message Vector {
    optional Complex start = 1
    [(complex) = { real: 0 imaginary: 0 }];
    required int64 length = 2;
}

See the Custom Options section for details.

Write your own kernel

James Malloy will show you how!

Better than null

A rich, delightful post from Joel Neely on an alternative to returning null to signal "not found" or "No": have the caller supply a callback with appropriate methods for condition signaling. In Java:

interface Found {
    void present(final String name);

    void absent(final String name);
}

class Phonebook {
    private final Set<String> names = new HashSet<>;

    void add(final String name) {
        names.add(name);
    }

    void remove(final String name) {
        names.remove(name);
    }

    void find(final String name, final Found found) {
        if (names.contains(name))
            found.present(name);
        else
            found.absent(name);
    }
}

class LookupPhone extends Phone implements Found {
    private final Phonebook phonebook;

    LookupPhone(final Phonebook phonebook) {
        this.phonebook = phonebook;
    }

    @Override
    public void present(final String name) {
        ringUp(name);
    }

    @Override
    public void absent(final String name) {
        complainIsUnknown(name);
    }

    void placeCall(final String name) {
        phonebook.find(name, this);
    }
}

(Keyed directly into the editor: apologies for typos.)

Bill Bejeck writes on Guava EventBus. This is very similar to the "MagicBus" code I wrote for a client while at ThoughtWorks in the mid-naughts. Essentially it is an in-process event bus respecting Java method signatures rather than a string-based topic space. You can write decoupled code in a pub-sub style within your own application.

In some weird sense it is an implementation of the COMEFROM instruction. When a "caller" publishes the right parameter list, subscribing methods are invoked as if called directly. Adding new subscribers changes the meaning of the call site. The difference is that control returns to the publisher after all subscribers run; COMEFROM is a complete transfer of control and does not multiplex.

Long live INTERCAL!