The Craftsmen Have it Right: Defining Software Quackery

I’ve fallen a bit behind on listening to some of my go-to programming podcasts of late (I will explain why shortly), but some of my friends were talking at dinner last week about a recent episode of .NET Rocks, featuring Alan Stevens. It presented the question, “Are you a craftsman?” and adopted a contrarian, or self-described “Devil’s Advocate” take on the movement. As far as episodes of that podcast go, this one was sort of guaranteed to be a lightning rod and the volume of comments seemed to bear that out, with Bob Martin himself weighing in on the subject.

I listened to this episode a week ago or so, and I’m not staring at the transcript, but here are some of the points I remember Alan making:

  • The craftsmanship movement in some places has turned into a monoculture — a way for the in-crowd to congratulate itself for getting it.
  • The wholesale comparison to guild culture and artisans is generally pretentious.
  • The movement should be inclusive and about building people up, rather than disparaging people.
  • There are lots of people out there not doing the “craftsmanship stuff” and still shipping working code, so who are we to judge?
  • Comparing software development to high-end, artistic craftsmanship, such as that of 30K guitars, devalues the latter**

(**I’ll just mention briefly that I consider this point to be absurd since this is shifting the goalposts completely from the “craft guild” comparison upon which the name is based — medieval craft guilds were responsible for making walls, shoes, and candles — not guitars for billionaire rock stars)

The comments for this episode contain a fair amount of disappointment and outright anger, but neither of those things was what I felt while listening, personally. Stevens is fairly engaging, and he had a number of pithy quotes from industry titans and authors at his disposal, citing, if memory serves, “The Pragmatic Programmer” and “The Mythical Man Month” as well as being fairly facile with the “Software Craftsmanship Manifesto.” This is clearly a well-read and knowledgeable industry professional as well as a practiced speaker, and he said a lot of things that I found reasonable, taken individually. But I couldn’t shake a vague sense of the surreal, like the dream sequences in the movie in Inception where Dali-like non-physics lurks at the periphery of your vision. It all seemed to make sense at first blush, but it was all wrong somehow.

I figured it out, and I’ll address that with a series of vignette sections here and hopefully tie them together somewhere this side of intolerable rambling.

On Quackery

One of the reasons I’ve been a bit negligent on my developer podcast listening is that I’ve become absolutely mesmerized by a podcast called “Quack Cast“, in which an infectious diseases doctor addresses alternative medicines. This man is ruthlessly empirical and relentlessly rational and even though the subject matter has virutally no overlap with anything I do, I find it fascinating. He provides ongoing rebuttal to all sorts of news about and advocacy for what he calls “SCAM” (supplements, complements, & alternative medicine), hoping to shine a light on how magic-based many of these approaches are.

Now, it isn’t my aim here to get into some kind of quasi-political fracas about whether prayer helps with surgery or whether aromatherapy cures cancer or whatever — if you have your beliefs, be welcome to them. But I will briefly describe one thing that’s so absolutely bat%&$# nuts that I don’t think it’s controversial for me to call it that: homeopathy. Now, if you’re like I was before hearing the podcast and then doing my own research, you probably don’t know exactly what homeopathy is. You probably think that it’s “home remedies” or “holistic treatment” or something else vague. It’s not. It’s much, much more ridiculous. It’s about 150 stops further down the crazy-train, firmly in the city of Quacksville.

You can read more here, but it’s a ‘theory’ of illness and cure that originated prior to the discovery of germs and the development of germ theory. Since that time, it has persisted largely and remarkably unchanged. The basic axioms are that “like cures like” and “the more you dilute something, the stronger it becomes.” So, for example, if you’re an end-stage alcoholic suffering from cirrhosis of the liver, the best medicine is tequila, but only if you cut it with water so many times that there’s no tequila left in your tequila. So, two wrongs (and one utter violation of the laws of physics as we know them) make not a right, but a nothing (i.e. giving water to someone with cirrhosis). And the fact that homeopathy is predicated upon diluting any actual effects out of existence is probably the reason it’s persisted while other historical, magical nonsense like alchemy, bloodletting, lobotomies, and exorcisms have gone extinct — homeopathy is relatively harmless because it’s all placebo, all the time (unless a practitioner makes a mistake and creates a nostrum that actually has some effect, which would be a problem). At the time this was proposed, it was as reasonable as any other contemporary approach to medicine, but in this day and age, it’s utter quackery.

Let’s come back to this later.

I’m OK, You’re OK

In software development, we start out as babes in the woods with an “I’m not OK, you’re OK” mentality. Everyone else knows what they’re doing and we’re lost. This is not healthy — we should be nurtured and not judged until we feel that we’re OK. Likewise, what separates good mentors from Expert Beginners is the adoption of an “I’m OK, you’re OK” attitude versus an “I’m OK, you’re not OK.” The latter is also not good. I write code and I’m OK. You write code and you’re OK. Maybe you’re more experienced than me, or maybe I’m more experienced than you. Either way, that’s OK. I’m OK and you’re OK.

At the end of the day, we’re all writing code that works, or at least trying to. And that’s OK. It’s OK if it doesn’t always work. We’re all trying. Maybe I test my code before checking it in. I’m OK. Maybe I at least compile it. And that’s OK. Maybe you don’t. And that’s OK. You’re OK. I’m OK and you’re OK. As long as we’re all trying, or, if not always trying, at least showing up, we’re all OK. There’s no sense casting aspersions or being critical.

This is the mature, healthy way to regard one another in the industry. Isn’t it? As long as we show up to work for the most part and sometimes write stuff that does stuff, who is anyone to judge? We’re all OK, and that’s OK.

Can Anyone Recommend an Apothecary?

If I’ve learned anything from reading all of the Game of Thrones books, it’s that the Middle Ages were awesome. Dragons and Whitewalkers and all that stuff. But what gets less play in those history textbooks is the rise of the merchant class during that time period. This is due in large part to the emergence of merchant and craft guilds. Craft guilds, specifically, were a fascinating and novel construct that greased the skids for the increased specialization of labor that would later explode in the Industrial Revolution (I think that happens in book 6 of Game of Thrones, but we’ll see when it comes out).

Craft guilds became professional homes for artisans of the time: stonemasons, cobblers, candle makers, bakers, and apothecaries — who knows, perhaps even homeopaths, though they would probably have been drawn and quartered for their suggested treatment of social diseases. The guilds had an arrangement with the towns in which they were situated. The only people in town that could practice the craft were guild members. In exchange, a basic level of quality was guaranteed. You could think of this as a largely benevolent cartel, though I kind of prefer to think of it as a labor union, but with more dragons.

Members of the guild entered as apprentices, where they learned at the feet of masters (and their family generally paid for this education). After completing a long apprenticeship, the aspiring guild member was allowed to practice the craft, for a wage as a journeyman, working with (for) other masters to ensure that knowledge cross-pollination occurred. With enough savings and proof of his “mastery” (which I think may be the origin of the word “masterpiece”) the journeyman could become a master and open his open business. The guild policed its own ranks for minimum quality, forcing its members to redo, pro bono, anything that they produced not up to snuff. The guild also acted as a unit against anyone selling knock-offs out of a trench-coat on the street corner. (The DVD makers’ guild was known in particular for this.) The self-policing guild offered a minimum standard for quality to the townsfolk and, in return, it was the only gig in town.

In practice, this meant a stamping out of impostors, charlatans, cheaters, and cranks. It meant that standards existed and that the general populace could depend, to some degree, on a predictable exchange of value. It also meant that there were accepted tools of the trade, processes and hoops through which to jump, internal political games to be played, and a decent amount of monoculture. And, it probably meant that the world progressed at the pace set by high achievers, if not geniuses. After all, the market economy had not yet been invented to reward wunderkinds, the outsized influencers, the lucky, and the radical innovators. Improvement happened, to be sure, but it happened at a pace approved by a council of masters that were ultimately men and men with egos.

So, this is probably the perfect metaphor for software development. Right? It seems like it’d be good to strive for a minimum level of quality — a set of standards, if you will. It seems like it’d be good if there was some kind of accreditation process, perhaps more accurate than a CS degree and less myopic than a certification, to demonstrate that someone was competent at software development. It seems like it’d be good for aspiring/new software developers to have a lot of one-on-one learning time with experienced developers that were really good at what they were doing. It seems like it’d be good for those initiates then to travel around some, broadening their experience and spreading their ideas. But, wait a second… we live in a very non-insular world, work in a global market economy, and, libertarians that we are, we’d sooner die than unionize. And writing software isn’t “craftsmanship,” the way that making candles, pies or shoes is.

Crap! We were doing so well there for a while, and there’s nothing more irritating than having to throw out your babies every time they dirty up their bath water. This seems to happen with all of my metaphors if I dissect them enough.

Software Quackery

Like medicine, artisanship, and even later 20th century pop psychology, software development has sort of lurched and hiccuped along in its (relatively short) history. Things that were widely done in the past have fallen out of favor. And, while a lot of our industry practices tend to be somewhat cyclical (preference for thin versus thick clients, for instance) some ideas do stick around. We make progress. We can stand, to some degree, on the shoulders of those who came before us and say that code at the GOTO level of abstraction does not scale and that shortening feedback loops and catching mistakes early are examples of preferable practices to their alternatives. We learn from our mistakes as individuals and as a collective. The bar inches higher. Or sometimes it stays stubbornly stagnant for a while and then makes a jump, the way that the field of medicine did with practices like hand-washing prior to surgery (except for homoepathic surgery to remove gangrenous limbs, in which case hand-washing is a strict no-no).

Movements emerge in fields, and they don’t emerge in a vacuum. They emerge in contrast to a prevailing trend or following a new, avant-garde idea. They are born out of words like “never again” as veterans of some practice look over the scars it has inflicted upon them. They become zeitgeists of the period before later being described as historic, or perhaps even quaint. In our field, I can think of some obvious examples. The Agile Manifesto and its subsequent movement springs to mind. As do the Software Craftsmanship, well, manifesto, and movement. Those are biggies, but there are plenty of others (everything on the desktop, no, wait, let’s move it to the browser, no, wait, let’s put it on people’s phones!)

So wherefore art thou, Software Craftsmanship? Well, I might suggest that Software Craftsmanship movement exists not in a vacuum and not to make those participating feel good about themselves after endless conferences, talks, and meetups filled with navel gazing. It’s not about condescension or creating “one true way.” It’s not about claiming that software is some kind of rarefied art form, nor is it about aesthetics or the “perfect code.” It’s not about saying that there’s one true MV* pattern to rule them all or that your unit tests need exactly one assert statement. Rather, at its core, I believe Software Craftsmanship is a simple refusal to tolerate Software Quackery any longer.

Once upon a time, back when PHP was new, we, as an industry, edited code on production servers, just as the founder of homeopathy thought that eating bark would cure malaria back in 18th century. However, a lot of time, study, experimentation and experience later, we came to the conclusion, as an industry, that hand editing code in production is a bad idea, just as medical science later came to understand why malaria occurred and how to prevent it. So now, it’s not pretentious nor is it beyond the pale for us, as an industry, to look upon software consultants editing code in production as quacks, anymore than it is to look upon medical practitioners handing out bark-water cocktails to treat malaria as quacks. We’re OK but you’re not OK if you’re doing that, and it’s OK for us to point out that you’re not OK.

It’s tempting to adopt a live and let live mentality and to be contrarian and say, “hey, we’re all trying to fight the disease, amirite, so let’s not be high and mighty,” but the fact of the matter is that treating malaria with bark and water is grossly negligent in this day and age, and it’s quackery of the highest order. So the Software Craftsmanship movement takes a page from the craft guilds of yore, and says, “you know, we should establish some minimum collective standards of competence, have some notion of internal quality policing, encourage aspiring members to learn at the hands of proven veterans, and develop the ability to say, ‘I’m sorry, but that’s just not good enough anymore.’” Yes, fine, if you ride the metaphor hard enough, it breaks down, but that’s really not the point. The point is that the movement is attempting to raise the bar from a world of barbaric medicine via unguents, spells, dances, and hitting people with rocks, to a world of medicine via the scientific method.

Back to the Podcast

As I mentioned earlier, I felt neither angry nor disappointed listening to Alan Stevens. I liked the conversation and at least parts of what he said. I picked up a few interesting quotes from authors and thinkers that I hadn’t hear previously, and his tone was relatively humble and disarmingly self-deprecating. And there was certainly a kind of populist, “don’t look down on the dark matter developers” vibe. So what was responsible for my surreal feeling? What was ‘wrong’ that I figured out? I could sum it up in a simple phrase, when it hit me: it felt like Stevens was pandering to what he thought was a low-brow audience.

What I mean is, here was a man, clearly knowledgeable about the industry, armed with impressive quotes and enough cachet to be asked to appear on .NET rocks, telling the listening population a very odd thing. Sure, he has a highfalutin Harvard doctorate like all those other doctors, but unlike them, he’s here to tell you that you don’t need any kind of fancy degree or medicine or machine to treat your own “carcinoma” or, as real people say, butt cancer — all you need is your own know-how, some kleenex, and a bucket of water. Or, in software terms, it’s okay if you ship crap, so long as you have a good attitude. I mean, we can’t all be expected to do a good job, can we? He even came out and said (paraprhased, though I remember the message very clearly) something along the lines of “sometimes mediocrity is good.” So, forget all of these fancy best practices and be proud of yourself if that 80,000 line classic ASP file you hand-edit in production manages not to kill anyone.

The underlying message of the show wasn’t any substantive indictment of the Software Craftsmanship movement, but an endorsement of Software Quackery. I mean, sure, there are good practices and bad practices, but let’s not get all high and mighty just because some doctors don’t wash their hands before operating on you — I mean, you’d rather have all of the colors of the competence rainbow than some surgeon hand-washing monoculture, right? The reference to Dreyfus really brought it full-weird-circle for me, though, because championing Software Quackery is generally the province of Expert Beginners — not Experts. So, I’m sorry, but I just don’t buy it. We can and should have some kind of minimum standard that isn’t a goody bag for just showing up. Please, join me in a polite refusal to tolerate Software Quackery — it just doesn’t cut it anymore.

  • Quagmire

    Great Post!

  • Andy Bailey

    Well chewing bark did actually prevent Malaria, as long as you were chewing bark from the right tree.
    Thoughtful post, I am also an advocate of minimum standards where the bar should be expected to rise as time goes on. I also enjoy the nurturing role of Mentor as it motivates me to ensure that what I know and my level of experience is up to date and of sufficient quality.
    I also learn a lot from those I mentor too, something that is often overlooked.

  • http://www.limulus.net/ Eric McCarthy

    So here’s the thing: Homeopaths consider themselves professionals too. They have organizations, conferences and journals where they publish papers. They even have certifications. Their craft has been around for over 200 years and they’ve developed tons of best practices through shared experience.

    Considering all that, can you really be certain that what Software Craftsmanship promotes isn’t also quackery?

    I’m not saying that’s the case. I know what Software Craftsmanship is, but I don’t keep up with the latest goings-on. However I haven’t gotten the impression that folks are reading and evaluating the studies that have been done or performing controlled experiments to test hypotheses themselves.

    If you haven’t seen Greg Wilson’s talk, “What We Actually Know About Software Development, and Why We Believe It’s True”, it’s worth the hour: http://vimeo.com/9270320

  • Brian

    I agree, the software profession needs to establish a set of core standards and best practices. And we need to do it now. When I put my EE hat on for a spell and then look back over the fence, software is a joke in terms of discipline and best practice. And this hurts me because at heart, I am a software guy and always will be.

    Sure, one can argue about the politics it breeds, but what about the politics that exist otherwise? Why, in this day and age, do I want to argue with someone over the use of source control? Why do I want to argue with someone about separation of concerns? Or, over the need to at the very least flowchart complicated aspects of a systems? Simply put, I dont. And I shouldnt have to consider it. Environments like this put responsible people in the position of either a) knowingly walking the road to failure or b) trying to play politics and “manage up”.

    In retrospect, I would have reached my current skill level about 5 years earlier if I had been properly mentored. But I wasn’t. I learned a lot on my own. I made lots of mistakes that amounted to real money being lost. I now find myself wholly inline with the “craftsmenship” movement, because I have essentially “reinvented” software development. I think this is true for a *lot* of developers. It shouldn’t be this way.

    We need to do better. New languages, IDE’s, and frameworks are not going to do it – ever. It starts with discipline, pride, and best practice forming the bedrock of our profession.

    IMHO, of course :-)

  • http://www.daedtech.com/blog Erik Dietrich

    Thanks — glad you liked!

  • http://www.daedtech.com/blog Erik Dietrich

    I think the gently rising minimum is really what I’m defending and advocating for, in the end. To your point, chewing certain bark had some efficacy and that was fine before the days of germ theory, but now we have inoculations and the like, and these should be the new minimum standard.

  • http://www.daedtech.com/blog Erik Dietrich

    I’d never seen that talk previously, but I watched and enjoyed it. When I was in grad school, I read a lot of white papers about software topics and saw that there were some studies done about the practice of developing software, but to the point here, not as many as you’d hope. I get what you’re saying about it being conceivable that Craftsmanship, based on a set of, say, beliefs, may also be software quackery.

    Here’s my take on that. First, I’ll say that I have, in fact, read white-papers and studies whose findings fall in line with what the Craftsmanship movement and, especially, Bob Martin, promote. Two examples come to mind off the top: smaller methods have been shown in studies to produce fewer bugs per LOC and code review cutting down on bugs (except Craftsmanship advocates ongoing review in the form of pair programming). So, while some of what the movement advocates is undoubtedly based exclusively on anecdotal evidence (which is certainly a prerequisite for quackery), parts of it are definitely based on empirical study of our craft, limited though the body of that study is.

    I’ll also say that our industry is young compared to, say, medicine. So, there’s a maturity spectrum, as I envision it. At the furthest end from maturity is what I’m calling quackery — Expert Beginners in random shops ruling the roost and refusing to allow source control usage because they don’t like it or something. Next furthest along is going to be the body of knowledge and set of standards endorsed by knowledgeable industry professionals. This is far from bullet proof, but it’s where Craftsmanship falls, IMO. You have a whole bunch of people with a proven track record for success sharing a common play book. Sure, it’s based at least partially on anecdotal accounts and personal experience but it’s better than nothing. The craft guilds managed to have standards for building things and fashioning products even though the scientific method hadn’t yet been developed at that time. Furthest along — where science based medicine is currently and where hopefully we’re going — is industry standards based on repeatable, peer reviewed studies.

    My take, anyway. It may well be that in 20 years some of the things we do today that are considered “Craftsmanship” may be regarded as quackery. And that’s fine. The bar should always be rising.

  • http://www.daedtech.com/blog Erik Dietrich

    Well said. The examples you cite here are perfect — particularly source control. I like the attitude of the Craftsmanship movement in this regard (and find Alan Stevens bemusing) for the simple reason that we need to stop having arguments about the basics if we’re going to move forward meaningfully. If doctors and medical scientists were still spending substantial amounts of time having the “germ theory versus ill humors and sin” debate in this day and age, it would be a colossal waste of time and effort.

  • Pingback: Making A Harvest Table – Remember this Tip ! | ideasandtools.com