It’s time to take another look at influential software engineering work

Yeah, really. Joel on Software was very well known in the software engineering community in the early 2000s and many of the posts are still frequently cited. Not only that, as founder of Fog Creek Software, Trello, and Stack Overflow (where he’s still CEO), Joel Spolsky continues to have a significant influence today. Looking back on some of his posts … well, let’s just say that they do not reflect today’s best practices for creating diverse, inclusive, high-performing teams.

And it’s not just Joel. Most classic software engineering work was done in an environment where almost all the loudest voices were guys and almost all of them were white — look at the signers of the Agile Manifesto, or the 14 guys (and no women) quoted along with me in an Economist article on bug finding a decade ago. Not only that, many of the loudest voices cut their teeth in environments like Microsoft (where Joel spent quite a few years, as did I) or the Linux world which … well, let’s just say that these environments were not always particularly friendly to women, people of color, trans and non-binary people, disabled people*, and those at the intersections of these and other dimensions of power.

Our understanding has come a long way since then.** There’s plenty of great research, well-understood best practices, and emerging techniques in areas like compassionate software and compassionate coding, ways of countering algorithmic oppression, gender HCI, feminist HCI, and post-colonial computing, collective intelligence, and other diversity-friendly software techniques.

But the old work is still out there, it’s still heavily cited, and the companies that were built using these approaches (and embedding the values that are implied) are still dominant in the software industry. When I give talks at conferences, a lot more people know about Stack Overflow and “Joel on Software” than gender HCI. So it’s not an exaggeration when April Wensel talks about how these articles have contributed — and continue to contribute — to the mess that the industry is in today. Doing something about it seems like a really good idea.

In the Twitter discussion April’s post kicked off, Fog Creek’s new CEO Anil Dash mentioned that they’re revisiting some of these older thoughts and ideas, for example in this post relooking at whiteboard interviews. That’s certainly a good thing (as is Fog Creek’s decision to put somebody like Anil who really understands the importance of diversity in charge). But there’s certainly more that could be done …

One of the big advantages that blogs and other online publications have over printed magazines and journals is that it’s often very easy to update them. So a good first step for authors whose past writing doesn’t reflect their current perspective is to update it — ideally with links to posts that better reflect what they think today, but at the very least to say “please don’t treat this as valid any more.” True, that’s a lot of updating for somebody as prolific as Joel. Then again, there’s a lot to be said for making an effort to clean up your past mistakes — and, for a company like Fog Creek, to signal that they’re trying to get beyond these ideas of the past. So this part seems like a no-brainer.

And just to re-emphasize something I mentioned earlier: it’s not just Joel. There are lots of other successful software engineers out there who should consider taking another look at their earlier work and inserting updates as appropriate.**

As well as Fog Creek taking another look at some of their old ideas, it’s even more important for them to get others’ perspectives. Some of the most important thinking is likely to come from people who would never even consider working at Fog Creek — or wouldn’t get hired if they applied.

This project is even more work than the simple updates described above, which is why my tweet very intentionally used the word “funded”. So many of us care about making the software world more inclusive and humane that it’s certainly possible that some people would contribute to the project without getting paid. But many don’t have the luxury of spare time to put on an interesting project; and, best practices these days are to pay people who do diversity and inclusion work (rather than asking them to do it for free as an unpaid second shift).

Yeah, I can see why. Fortunately, there are plenty of other people who would also have more evolved takes — for example (and in no particular order) Jill Dimond of Sassafrass Tech Collective, @Shaft of Venture for America, Kimberly Bryant of Black Girls Code, Sarah Mei of Salesforce, Danilo Campos of Vaya Consulting, Shireen Mitchell of Stop Online Violence Against Women, Kaliya-IdentityWoman of the Internet Identity Workshop, Stephanie Lampkin of Blendoor, Marco Zehe of Mozilla, Christie Koehler of Authentic Engine, EricaJoy of Patreon, Leigh Honeywell of the ACLU, Ari Schlesinger of Georgia Tech, Tammarrian Rogers of Snap, Kelly Ireland of O.school, Wayne Sutton of TechInclusion, and Gayle Laakmann McDowell of Cracking the Coding Interview. All of them would have some really interesting things to say … and they’re just the tip of the iceberg. [Other suggestions welcome!]

Also in the same way that “it’s not just Joel”, it’s not just Fog Creek. Many other software companies would also benefit hugely from taking a look at their past work — and the past work that’s influenced them — through different eyes. In the software engineering space, Stack Overflow and Microsoft are obvious examples; so are Atlassian, GitHub and its competitors, and ThoughtWorks (one of the few large software engineering shops that’s long been conscious of diversity).

So hopefully people will decide to invest in making this happen both at the individual and corporate level. It’s a great opportunity to clear up a lot of destructive misconceptions — and help technology become more inclusive.

Thanks to April Wensel and Laverne Masaki for taking a quick look at an earlier version of this post!

* see s.e. smith’s Why I Say ‘Disabled Person’ Instead of ‘Person With Disabilities ' on Rewire for why I’m using “identity-first language” here.

** Techniques for “Supporting diversity with a new approach to software” has many relevant links. And of course there were people who understood it back in the day as well; Justine Cassell’s 1998 paper on a feminist approach to software design is a good example.

*** Including me! Admittedly I’m nowhere near as prolific or widely-read as Joel, but there certainly are some things I got wrong in posts or presentations that still sometimes get attention; now’s as good a time as any to start fixing them. So over the next few months I’m going to look at some of my own past work and updating where it makes sense. Hopefully others will do the same!