When approaching a new project, Elsevier technical lead Justin Florentine sees Oracle's Java 8 — an update to the programming language first made available in 2014 — as a red flag.
"Older versions of a software are generally a bad sign, to be honest," said Florentine, a 20-year Java veteran who joined the information analytics company in 2016. "We treat that as a code smell, it's a clue that leads us to problems."
Java 8 is among the most stable and backward-compatible releases of the software. Eight in 10 Java developers used it in production for their main applications last year, according to a survey of 10,500 Java users.
But the platform's reliability could change, in light of Oracle halting free public updates for commercial users in January, with additional support available through a paid plan.
"The problem is that, at some point, someone will force your hand and you'll have to update."
Technical lead at Elsevier
As tech managers evaluate the need for updates, experts say a mix of empathy and common sense ought to handle the tension between retaining older, more stable technologies or embracing new, flashier frameworks.
With older tech platforms, complications can arise as resources become scarce. Florentine cites one hiccup during a system update to Java 9. Developers had used a library owned by Oracle that was not part of publicly available resources.
"The problem is that, at some point, someone will force your hand and you'll have to update," Florentine told CIO Dive. "And by that I mean something far more tangible, like a security vulnerability."
Often, the decision to not update systems has to do with the lack of human resources and time to deploy updates to mission-critical systems. Think of how long companies took to update out of Windows XP, for example.
"It's in IT's interest to stabilize that tech as much as possible because it runs the business and is critical to the mission," said Jeffrey Hammond, VP and principal analyst at Forrester, in an interview with CIO Dive.
Weighing the cost
Amid a crowded field of new and enticing technology, CIOs must rely on their ability to build strong teams, listening to their insight for guidance, said Vicki Boykis, senior manager of data science and engineering at Captech Ventures.
Having seasoned technologists aboard helps add nuance to the decision making process, since they "will look at a technology and say: 'maybe this doesn't make sense, let's think about it,'" Boykis said.
"You can have 10 years of experience or 10 times one year of experience where you just make the same mistake over and over again," Boykis said. "This is hard to hire for, but it's about hiring the person with 10 years of experience."
"If you’re going to do something new the benefits should outweigh the risks."
VP and principal analyst at Forrester
The world's IT already runs on less-than-flashy technology, like Excel, Java 8, and Sharepoint, and tech professionals should be empathetic of that reality, Boykis said, in a May 10 blog post.
Any retool of technology, Hammond said, must prove to be worth the cost. Newer platforms can make their case by offering more stability than their predecessors.
"If you're going to do something new the benefits should outweigh the risks," Hammond said.
As decision makers aspire to build life-changing technology, managers should put in place measures that reward the better and more appropriate resource for each case.
"Defect density is one measure," Hammond said. "If you're writing in Java your systems should be pretty bug free, since there's been a lot of things knocked out of systems that might not be the case with newer languages."
Companies can recruit sufficient talent for Java 8, and existing databases are robust, Florentine said. More trendy languages like R — heavily used for things like statistical analysis and data science— lack the infrastructure Java has.