[Software] IBM Developer Lead: What Software Does Next

[_Happy boy]

Software changes constantly. There is an ever-shifting and constantly evolving demand for more software application development as users’ use cases continue to change. This perpetual dynamism comes about because industries and world markets change, commercial relationships alter, workplace methods evolve, new devices come to market and enterprise and consumer demands cross chasms and become markedly different.

Quite apart from all those reasons, software platforms, protocols and processes themselves change, which (you guessed it) creates software change. Floppy disks become terabyte servers, ‘old-fashioned’ approaches to networks become instances of cloud computing, traditional programming models get shaken up with new maverick ideas… and all manner of legacy software gets slowly consigned to the ‘where are they now’ file as we find new ways of creating apps.

With the current renaissance in Artificial Intelligence (AI) that drives automation (bots and more) and autonomous computing, the growth of big data analytics and the future promise of quantum power, things show no signs of slowing down either.

IBM Garage: software spanners, sockets & soda
Working to try and anticipate what comes next, how to harness current programming methodologies and build software for tomorrow and the day after is Dr Holly Cummins in her role as worldwide development community practice lead for IBM Garage. As detailed here, IBM’s Garage initiative is a combination of consultancy services, shortcuts, best-practice templates, reference architectures and the chance to work with coders in the code garage, getting greasy, or at least virtually greasy, especially in times of lockdown.

Having seen her fair share of the good, bad and perhaps even ugly side of software application development, Cummins says she has led enterprise-grade software projects to count fish, help a blind athlete run ultra-marathons in the desert solo, improve health care for the elderly and change how city parking works. She came to ‘love’ computers from an early age because her father had a PC at home, even before there were IBM PCs. It was a CP/M machine – now lost in programming history - but very applicable to use for learning the BASIC programming language and playing a few games.

MORE FOR YOU
Health-Tech’s Tough Pill To Swallow Is ‘Data Ingestion’
An AI Engineer Walks Into A Data Shop...
After Localization, Software Needs ‘Countrification’
But Cummins (who often goes by Dr Holly) didn’t study computer science at university. She thinks that this is one of the brilliant things about computing as a career and insists that it is truly accessible and people can come into it from all backgrounds, which enriches the whole field. She says that there’s a debate raging at the moment about whether you need a computer science degree, because some of the best developers she has worked with have English majors. 

Before DevOps, we still did dev & ops
Given the fact that IBM Rational events were talking about conjoined developer & operations techniques some 20-years before the term DevOps was coined, does Cummins think that the market has simply productized and po[CENSORED]rized an approach that IBM was already championing?

“The reason DevOps, as an idea, is being regularly reinvented is that ‘true DevOps’ is really hard. Rather than doing it, we often just misuse the DevOps term for other things. We often say DevOps when what we mean is Continuous Integration and Continuous Deployment (CI/CD) – and even with that, the integration and delivery isn’t continuous at all, but happening at intervals, like every six months, so it’s actually just I/D i.e. Integration & Deployment. True DevOps, by which I mean weaving together developers and operations in the first place needs to overcome a lot of barriers; people are the hardest part of computers. Making sure the people are aligned doesn’t happen easily, which is why we’re still having this conversation,” she said.

Cummins speaks of her ‘running an accelerator’ reality theory that advocates expecting the unexpected. This is a concept in two parts: frontloading the value and frontloading the risk. She explains that we’re often in the habit of doing things the way we’ve always done them... and in an accelerator context, but that just doesn’t work. We need to first identify the bit that’s really hard and even though that’s the most uncomfortable part, it’s the bit we need to do first. This comes down to the fact that organizations today need to maximize value in time-scales of one week, rather than six months.

The key is to focus on the people and the problem that they have – focusing on the problem, instead of the technology and the solution. Technologists don’t always want to do this, but ultimately, is what will unlock innovation.

“Here’s an example. A few years ago, chatbots were really po[CENSORED]r, so a client wanted to do a chatbot to allow factory workers to reset their password when they got locked out of the corporate IT system. A chatbot was pleasingly similar to the employees’ current way of operating i.e. doing it through their phone - but the most similar option isn’t always the right one. We had to unpeel the layers of an onion to get to the real problem. The reason users kept locking themselves out and required so many resets was because they had really heavy gloves on and only had three attempts to enter their password on a little handheld device. So the actual solution came down to making sure that they had six attempts, rather than three. The most valuable solution wasn’t a big tech investment but a modification of an existing process,” explained Dr Holly.

Method in the mechanics: inside the IBM Garage
In terms of what’s going on inside IBM Garage right now, Cummins and team say that - despite the upheavals of 2020 - there are some home truths surfacing. As a self-confessed ‘methods’ geek, she says that one thing the Garage team has witnessed over the last half decade is the fact that an implementation method that started out for small businesses getting to the cloud and adopting AI, actually makes sense for large businesses too.

“As a transformation accelerator, Garage is really good at getting people aligned on the problem we want to solve. Often, clients will come to us with a big pile of data and just say: let’s use this. But Garage lets us step back collectively and figure out the questions we want to answer – what will really make the biggest difference to modernize their business - before diving into the data,” she said.

Dr Holly is sanguine, measured and generally upbeat about what comes next in IT and is aware of some of the hype that’s out there in terms of emerging (and fast growth) technologies such as serverless, Robotic Process Automation (RPA), AI, orchestration and observability. Of these, she points to RPA as a point of interest.

“‘[In terms of RPA], the whole point of computers is that they take things we don’t like doing and do them for us. Everything going on with digitalization and automation is liberating the human spirit to work on more interesting things. Similar to this, AI is also about understanding our limits as people and what computers can do that we can’t, but also remembering what we can do as emotional empathy-aware humans that computers have no hope of doing. We want to get to a point where AI does what it is good at and people do what they are good at,” said Cummins.

She also points out the eco-benefit climate change potential that we can potentially achieve by using serverless computing techniques. If we can strip out a bunch of stuff from what we’re doing to compute and move towards more managed infrastructure by pulling resources that are more efficient for everyone, both financially and ecologically, she thinks that’s great.

Being brutally realistic and honest on the subject of orchestration and observability (which we generally use to talk about cloud asset insight and management), she hopes that we soon get to a point where orchestration and observability are a) taken more seriously and b) more baked-in almost as computing utility, so that we can get on with higher level concerns.

Living the quantum computing dream
“Another are to mention is quantum. When I did a PhD in quantum computing, we didn’t actually know if there would ever be a quantum computer. But now, you can just go to the cloud and use a quantum computer and that’s incredible. I think we’re going to continue to see this incredible progress. The thing that’s really interesting is that we don’t know yet what problems it should be solving – we just know there are unsolvable problems, so we don’t even formulate the question. Now we have to go back and find the new questions to see what we can solve,” said Dr Holly.

She also speaks about the perils of micro-optimization, over-containerization, the frustration of over-orchestration and (a very lovely term) kubesprawl in the world of Kubernetes (let’s call it kubesprawlerization for consistency), so does she think we have become too granular in computing?

“I think the issue there isn’t necessarily granularity but focusing on the shiny obvious problem rather than the key problem we need to solve. We too narrowly articulate a problem and lose sight of what would have the most impact. We wind up spending too much time thinking about our orchestration, for example, because it seems like we’re adding value. You shave 20% performance off some process and you think you’re a hero, but you step back and realize that you do that process every three months. Maybe there’s another process you do multiple times a day and you’d see huge savings by focusing there instead – even with a smaller optimization,” said Cummins.

Looking ahead to whenever the next big thing is coming, Dr Holly is positive about open source and insists that it is the ‘bedrock’ of modern computing. After all, it’s powering all of the most progressive tech of our time, including quantum, AI and cloud.

At the risk of being a bit philosophical, she suggests that open source also shows something inspiring for us as humans i.e. it shows the best and most successful things happen if we bridge our differences and collaborate together for the common good. Pointing out that companies are now adopting an open hybrid cloud approach, she reminds us that developing open source skills is just common sense – they’re more transferrable across the developer community and ecosystem and across different companies. 
“Cultivating open source skills makes us better prepared to work together to solve the big problems. But, sometimes (probably most times) you never know the next big thing until you are looking back in the rearview mirror. At the time it seems like incremental improvements, but when you look back, you realize that there was a really huge change,” she concludes.

It’s easy to think of IBM as that ‘old’ technology company. It’s been around a long time, some of its execs still wear dark suits at its annual user conferences (not that we’ve seen one of those in a while) and it adopts a fairly serious matt-color-scheme approach to most of what it presents whenever it goes on public show (whether that’s conference stands, mainframe housing units or promotional T-shirts). It may take a while for IBM to completely change those spots. Perhaps it won’t. Maybe it’s just not in its nature.

Of course, there will be naysayers in every market for every product, IBM won’t be trying to adopt a new Hi-Viz orange and lime green livery for its brand reinvention any time soon. So we need to look inside divisions like IBM Garage to see where real innovation is still breaking new ground and ask purists to the cause like Dr Holly how the cogs are really moving on the inside. Big Blue is still blue, but it’s still big too.