CAMBRIDGE, Mass. – With the back-to-back keynotes all set, afternoon breakout sessions at the ReDev B0st0n Conference commenced with a very popular one to start.
Long-time PTC engineer Bruce Hulse had a packed house for his talk, Developing for the IoT Space. The Internet of Things (IoT) has been talked about for quite a while, with a good deal of hype about what it will look like. The buzzword is not new, but it is very much emerging right in front of us.
Much of what has been said and written about IoT has focused on what the end devices will look like and the kind of intelligence built into them. From knowing when you need to re-stock your refrigerator to a toaster that never burns your toast, the promises have been many and varied. To that end, though, Hulse got everything going by talking about what IoT applies to: what we make and the tools that help us make what we make.
It is the latter that may ultimately be of more interest – or at least, more impact.
Forget about the consumer novelties – some of which don’t necessarily strike everyone as overwhelming positives. Imagine if you and your company could use the data that is available on your production devices to keep production moving, proactively see potential upcoming failures and schedule maintenance before the production line is forced down for some precious time because something finally wore down and broke entirely.
There are many possibilities regarding the data, including transports for it (GPIO, I2C, SPI, RS485) and places it may reside (on the device, on the cloud, or a local host machine). There is also a wide skill set that comes into play, and at one point he had a big list – suffice it to say there is something for everyone with IoT.
Next up was Ryan Scharer from PatientPing talking about Kotlin, a relatively new language that is very Java-like. With Google giving the language their blessing by making it part of the Android ecosystem, the language clearly has a future. It also helps that Spring 5 will add support.
Kotlin is seen as a drop-in replacement for Java, and also an attempt to take Java and improve it. A few key areas Scharer noted were immutability, methods that are final, and null values. He gave a few examples and also noted that while it cannot mix with Java in source, it can operate with Java and JavaScript code. Anyone who knows Java should pick this up quickly.
When that was all done, it was on to From “Big” Data to “Big Data”, a presentation by Toast President and Co-founder Steve Fredette and cloud architect Martin Kressirer. Fredette started by introducing the company, which recently snagged $101 million in funding, and said they want to be the Salesforce of the restaurant business. They feel the restaurant business is currently a low-tech industry, but that is bound to change. To that end, Fredette noted the success of Domino’s, whose stock jumped after a big investment in technology, and said the company now views themselves as a technology company that sells pizza.
Fredette noted that there are many areas in which the company works with restaurants, and still much room for growth. Notably, a lot of the data they have access to is thrown away, but ultimately that may change given that it can serve useful purposes such as ordering, cooking and delivery, analyzing labor and inventory, and marketing and loyalty will likely grow as well. I first became aware of them from a loyalty program they help a coffee shop administer, and as more restaurants adopt such programs – many of their retail counterparts are adding them – there will be one more avenue for Toast to do business with restaurants.
The challenges are steep and never-ending, though, and they illustrated this with their experience. For starters, restaurants are a bigger data source than retail, easily determined just from the reality that we generally eat three times a day but almost never shop even once a day. The company didn’t always have the APIs and other technologies they do, but failures like an accidental distributed denial of service (DDOS) and seeing just how much data is generated taught them to get there.
Kressirer took attendees through a couple of times where their databases have been under a big load and have generated enough data that they would run out of space at one time in about four moths, then later a little longer than that. Every time they thought they had solved any kind of space issue, another one was right around the corner.
Fredette went on to talk further about the potential by noting that there is a long-term trend of eating out more, as well as a reversal in the farming population. At one time, only a small percentage of Americans did not work on farms, but now only a tiny percentage do. Additionally, Amazon buying Whole Foods Market goes with this. Aided by their recent funding, the company is likely to surpass 100 engineers on their staff before long.
The day was then brought to a close in fine fashion by Julie Yoo from Kyruus, who gave a great talk entitled Doctors Are Not Airplanes: Solving Capacity Optimization Problems in Healthcare. Right away, she grabbed our attention by noting that doctor’s offices are at only about 60 percent capacity utilization, and that comes after controlling for time they must spend on administrative and other tasks that do not fall under the umbrella of time with patients. We wait an average of 4-12 weeks to see a doctor, and Boston is the worst in the country at that.
All of this, plus the fact that healthcare historically spends less on IT than other industries, points to potential for technology to make a big difference. Right now, so much happens offline; we all know how you have to call a doctor’s office to schedule, modify or cancel an appointment, just for starters. Specialization doesn’t help, as evidenced by the example of how there are 20 different types of cardiologists and seven claims codes for atrial fibrillation.
Not only is there not so much spent on IT in healthcare, but what is presently considered the state of the art is not very good. Search terms need to be as precise as possible, and commercial search doesn’t work well at all. In many cases, a patient can’t search for a male or female doctor, in a certain language or insurance acceptance.
There is some good reason for healthcare being slow to adopt IT. Security is a very big deal in this space, especially with HIPAA. Yoo compared the current state of digital health to 1997 for digital computers, when the Internet was just coming of age. We can only hope that digital health has the same success we have seen in the computer industry since then.
With that, a great day at the ReDev Conference concluded. Many headed upstairs to partake in the networking reception and career fair that followed.