Proprietary software use plummeting? Well, maybe not quite

According to Red Hat’s new The State of Enterprise Open Source 2020 report, most everyone (95%) believes open source is important, and 77% expect to increase their use of open source over the next year. This seems reasonable, since nearly all (up to 90%, according to Sonatype) software includes open source components. Yes, that’s right: You might think you’re using a proprietary database/operating system/whatever, but much of the code within is almost certainly open source.

Those are the facts.

What’s a little harder to decipher in the report is the apparent decline in proprietary software. It may be that enterprises are buying less proprietary software, but it’s doubtful they are writing less, and that’s still the biggest chunk of software written.

SEE: 10 ways to prevent developer burnout (free PDF) (TechRepublic)

Opening up the enterprise

Over a decade ago Red Hat CEO Jim Whitehurst issued a clarion call for user-driven open source development: “Ultimately, for open source to provide value to all of our customers worldwide, we need to get our customers not only as users of open source products but truly engaged in open source and taking part in the development community.” This has happened to a small degree. Capital One, Lyft, and others have actively licensed software under open source licenses, and still more companies have participated in projects like Kubernetes.

Yet this remains the exception, not the rule.

Most companies are net users of open source, but not contributors thereto. Talking with a few large enterprises this past week, each faces a major hurdle in justifying why code should be open. (Of course, as an analyst friend recently related, that really should be flipped–we should have to justify why something should be closed, not open, but as an industry we’re not there yet.) Most enterprise software offers no compelling competitive differentiation (does your employee directory app really distinguish you? Or even the CMS you use for document management?), yet virtually all of it sits in private repositories.

And most of the world’s software is written for use not sale, same as it ever was.

“Proprietary” means what, exactly?

This brings us back to Red Hat’s survey. In it Red Hat finds, “The use of proprietary software is plummeting. Last year, our respondents indicated that about half (55%) of the software they used was proprietary. This year, it’s 42%. Two years from now, they say proprietary software will be down to 32% of their software stacks.” Here’s what it looks like in pretty charts:

I’m an open source guy. I’d love this to be true. But it’s not. Dig into any enterprise’s GitHub or GitLab private repositories, and most of the software they’re writing remains their own proprietary code. Are they using copious quantities of open source? Absolutely. But they’re simply not able to do everything they need with open source. Instead, they’re reinventing wheels that every other enterprise is building, with precious little sharing between them.

Those are the facts, too. But I don’t like these facts quite as much.

Perhaps what Red Hat was getting at with its survey is that organizations are buying less proprietary software. I suspect that’s true, as companies move off old-school databases for more modern, open source options like MariaDB, MongoDB, etc., or shift to other open source infrastructure. That all makes sense.

According to that same Red Hat survey, just 69% of legacy applications are expected to be left “as is,” while the remainder get modernized (most) or decommissioned (small minority). Perhaps as companies look to do this, they should also be looking to modernize how they build (and release) software. Using more open source is great. Contributing more open source is better.

Disclosure: I work for AWS, but nothing herein relates to my employment there.

Also see