On professionalism in software | Hacker News Reader

On professionalism in software

jakewins.com

52 pointsidlewan19 comments

I can agree, and I can disagree. I definitely wouldn't compare software engineering to medicine; rather, I see a more apt analogy in, actually, _engineering_ – whereas designing a firearm is just as noble and ethical

Software is eating the world, good and bad, and our role as a profession is to help make the transition of the entire society, altruistic and greedy, charitable and fraudulent, to the highly automated mode of being. After this, it's everyone's individual karma vector's job to guide us in more or less ethical way of conducting ourself professionally.

On your first point, I agree, the medical comparison did not come across as well as I hoped. The intent with "branching out" past engineering in the text was to highlight Software Engineering as a profession with ethical responsibilities. Engineering is a closer comparison - but medicine or journalism is much more mature in how they think about ethics, in my opinion. Engineering disciplines are apt to think of the toolmakers dilemma as solved by delegating responsibility to the user.

Which gets to your second paragraph which, while I really like how you've worded it, I disagree in the general case. Yes, we should build technology that is unbiased and pliable. But I think the example of TCP/IP or HTTPS illustrate that that is too simplistic. We are responsible for building technology that answers to its users, and highlights or removes the ability for users to be abused when interacting with it.

This can certainly be boiled to “when engineering a product, be thoughtful of how it can be (ab)used – if it can be, it will be”. Which is obviously a great advice from any perspective.

are you familiar with Michael O Church? https://michaelochurch.wordpress.com/2012/11/18/programmers-...

> We, as an industry, need to stop seeing users as a resource from which to extract value.

I mean, yes, but... at the end of the day this is a capitalist society, and if you want to build, say, Google, you do not get your funding by convincing a central planner that this will benefit people, you get your funding by convincing an investor that you can make them rich.

The question is how can we preserve the best interests of our users while creating profitable businesses, which is a hard problem that I do not know the solution to.

capitalist society is not a static - we can and must change this to benefit people

Ok, but what does that mean? Change how?

There's a zillion alternatives to capitalism. Those of us whose imagination got sucked out of our heads, unfortunately seem incapable of exiting a framework like neoliberalism's "capitalism vs. Soviet Union". As if there's nothing to drink beyond Coke vs Pepsi.

> stop seeing yourself as a lone individual developer

Sorry for a gloomy vision but:

1) the 'corruption' of the modern software business is driven by business and not technical leaders;

2) even vocal and principled engineers have kids to feed and even those without such commitments don't leave free massages in the office easily;

Those of us working in larger organisations are just that when it comes to fighting for principles - hopelessly lonely.

I was somewhat thinking about this when I wrote that sentence actually, and it's a topic that needs its own post at the very least.

That said, and with the caveat that I'm aware of the vast amounts of issues around this, I will simply say that others have thought about this problem before us. http://en.wikipedia.org/wiki/Portal:Organized_Labour

I think organised labour sounds too 19th century to most people in the business. None of my colleagues that I surveyed in my previous (big) company was supportive of those. Trade unions of old used to fight for better working conditions and what are they to fight for in the likes of Google? Better phones for Christmas? That may be an argument until for one reason or another you run foul of your boss or some awkward company policy and then you're completely alone. I guess though today's culture of individualism and the myth of meritocracy in tech companies work very well against any idea of organised labour.

> Trade unions of old used to fight for better working conditions and what are they to fight for in the likes of Google? Better phones for Christmas?

Something for software engineers at that level would, rather than being a trade union, probably resemble the professional associations doctors, lawyers or "real" engineers have. Like the ACM or IEEE except... you know.

> I think organised labour sounds too 19th century to most people in the business.

Just wanted to note that this is an eccentricity specific to the US.

Now is it? I actually live in Ireland and most folks around me think exactly the same. The local trade union [0] is on its deathbed and I haven't heard of a single IT multinational with any form of employee representation.

[0] http://www.siptu.ie/

Those companies start somewhere and these days are certainly driven by technical folks. Even Google was a small team once and yet the business model adopted was to harvest user data in order to target ads. Twitter took a similar route. I think it's disingenuous to attempt to separate 'business' from 'technical leaders', as if the latter have no involvement in the former.

(edited: see rest of thread)

> yet the business model now is to harvest user data in order to target ads

This was their original business model. Previous to that they had none and were pretty much considered doomed.

I edited to reflect that the business model was adopted. It doesn't alter my point.

Professionalism in software is a weird term. I think this essay would've better been named "On ethics in software engineering".

From a purely professional standpoint, we should've long ago settled on a few languages to do our tasks, and focused on exploring how to write and manage decade-long projects in those cases, or at least how to rapidly develop and shift codebases to meet the new year's business needs. We've done neither consistently.

We should've come up with some kind of licensing and liability system for work done, similar to the Professional Engineering license that is required for building a bridge or skyscraper. This in turn requires that we should've come up with a common nomenclature for solving business problems--we tried this with design patterns, but in a further fit of snowflakeism this approach was unfairly derided and made fun of by our entire industry, mostly because we let it get out of hand.

We should've come up with some kind of way of specifying and defining projects, so that we can give an honest quote ahead of time and give clients a way of explaining what they want built and removing the "oh but I guess that's not what I wanted" entitlement many seem to have because they (rightly) don't respect our professional expertise.

We should've long ago bridged the gap between the academic/theoretical practice of computer science and the professional issues of software engineering and the performance issues of computer engineering, but instead we decided that internships and coops (though not always by that name) were the "real way" to learn about software development.

We should've done a great many things, but we didn't, and now? Now it's probably too late to turn things back. So, in the mean time, let's make hay while the sun shines, and pass the buck to the next generation of poor hapless bastards.

Oh, and as for the users? I've got a rather unpopular opinion on that:

They've accepted, encouraged, and rewarded at every term their own enslavement and disrespect.

They've continually decided to ignore how the magic boxes work, decided to embrace the shiny over the functional, and in general cast us as the Morlocks and them as the Eloi.

So, we might as well exploit them---they neither notice nor want more.

EDIT:

And the downvotes, as expected.

Until we build software like we build bridges, we shouldn't faff about with professionalism. We need to stop coming up with language-of-the-quarters and code spikes and other stuff. And yes, that means that things aren't magic and fun and whatever, and that means that things aren't interesting, and that means that maybe we don't use the exact best tool for the job, but that's practical engineering. It's boring as hell but dependable and conservative.

Until our users realize that "hey, maybe I shouldn't spew all of my personal data to complete randos on the internets", they don't deserve to be treated as anything other than a monetizable resource. Until app users are willing to pay more than a buck for their mobile app, we should avoid helping them and instead maximize our revenue. They don't respect us.

Look, there are edge cases in certain types of bespoke software work, certain good clients and customers that are informed and intelligent, but the most of the sixth of the world that is on, say, Facebook, can neither imagine nor articulate the freedoms they forfeit (for themselves and others!) every day.

We do build software like we build bridges. Over budget, behind schedule, and with worrisome issues left hanging over our heads.

Reply on news.ycombinator.com