Sorry, Moxie. Blaming Agile for software stagnation puts the wrong villain in the wrong play

Innovation dead for twenty years? Tell that to 2004

Opinion Want a good time on stage, but you’re not a performance artist? Surprisingly easy. Fill a hall with an audience of your peers, tell them the world’s gone to hell in a handcart, then that they’re the only ones who can fix it. You’ll feel the love. Guaranteed.

This is precisely what security superstar Moxie Marlinspike did in his recent Black Hat keynote. His thesis was stark. Over the last twenty years, magic has gone out of software development, innovation has stalled and programmers no longer understand what they’re doing. The reason? The stifling straitjacket of Agile.

The cure? The full-stack deep appreciation of software engineering exemplified by security researchers. You are the wizards, the last bastion of true magic, Marlinspike told Black Hat. We can save the world!

A great story told by a true hero of security. Co-author of the Signal Protocol and prime mover of the Signal Messenger app, he also had the political chops to get that protocol into Meta, Google and Microsoft. It’s just a shame the story doesn’t stand up as well as his end-to-end encryption.

Let’s start with the twenty year innovation drought in software. Services delivered by browser were just getting going: the arrival of Google Maps in 2005 stopped work in offices around the world as people clustered around monitors to look. It looked magical. It’s easy to think that nothing much has changed since then, but only if you ignore the smartphone ecosystems which were to come. Also in the the future back then were AWS, Twitter, Docker, that whole microarchitecture thing, and all the infrastructure that supported the greatest expansion of connected high performance computing across the globe.

Innovation doesn’t equate to revolution. Evolution couldn’t happen without it, and we have evolved so very far since 2004. It stands to reason, therefore, that Agile hasn’t actually killed it. No engineering methodology can kill innovation: many can be mismanaged to dissuade creative thinking, but that's mismanagement. All the sins Marlinspike ascribes to Agile - siloing, big picture blindness, suppression of skills acquisition - are explicitly addressed by devops, which is entirely compatible with Agile.

Devops can be a car crash, but car crashes are rarely the fault of the car.

Moreover, it’s easy to forget how unstable a lot of software was twenty years ago. Crashes, freezes, dismal drivers were less common then than they had been in the previous decade - let’s not talk about the 80s - but far more common than now. That CrowdStrike BSOD was the first many adult humans had ever seen, adult humans whose use of IT is far more constant and intense than anyone was experiencing in 2004. Sending an octogenarian relative a new Chromebook when their old one breaks, and they can get all their stuff back in a few minutes? Try doing that with 2004 tech.

So if innovation isn’t locked in the frozen meat store of Agile, why does it feel to Marlinspike and many other practitioners of exquisite expertise that it is? When you’re closely bound into something, you miss the incremental evolution. Your skill set, no matter how superb, will fit differently. There’s still a need for those good old low-level clever programming skills, but let’s not forget how much time we’ve spent cleaning up some of the messes they created. The fact is, if you are an innovator or work for a company that sees innovation as its rocket drive, the tools for being innovative and getting it out there are hugely better than they were twenty years ago. All the old tools are still there, just with twenty years go new ones to use if you want to.

None of that will help if you’re still trying to fix old problems in the old ways. That’s been done. The new ways have created new problems, and that’s where new thinking is most needed. Look at where digital technology is hurting when it should be helping and get smart about it. Things that have gone backwards - accessibility, user autonomy, transparency of what and where and how your data is being used, tools of trust and curation.

All rich areas where only software can make things better, and only software that a disruptor can deliver. You want to recreate that Google Maps moment of twenty years ago? Create a Google Maps for the digital environment, something that anyone can look at and use that describes where their digital self is and what’s happening to it. Create a magic wand that can take the horrifying mess of all those dark patterned corporate websites and services and remaps their functions to something that dehumanizes the experience.

There remain countless opportunities for software innovation that is revolutionary, not just evolutionary. That this innovation isn’t happening isn’t because corporates are using Agile, it’s because the sort of innovation that’s big enough to be noticed is not the sort of innovation that a lot of tech corporates want. Especially the big ones.

Telling people they’re great and can save the world from an easily identified evil - that’s great. It’s much harder work to say to your audience that they are magicians and high priests of their domain, and things are going wrong, the answer doesn’t lie in the past.

Looking at the world through fresh eyes and taking control of the new ways is the only way to stay relevant. There’s even a little phrase that sums it all up - be agile. ®

More about

TIP US OFF

Send us news


Other stories you might like