A myth about COBOL, the Common Business-Oriented Language invented in 1959, is that large organizations use it only because it’s too tedious to rewrite their programs in something modern and because nobody young knows how to move their employers away from it.

Reality begs to differ. Large organizations use COBOL because it’s incredibly efficient at high-end transaction processing applications in ways that no newer language can match; the programs operating today are largely based on COBOL updates from 2002 or, at worst, 1985, and learning how to write its code can be mastered in a few weeks by any competent C or Java developer.

COBOL remains popular for IBM’s z series of mainframes and for non-IBM distributed systems, too. “It’s solving a very important problem in a good way, and people continue to use it,” noted Kevin Stoodley, IBM fellow and CTO for z Systems software.

The language is steeped in computer history. COBOL began in 1959 through an industry standards committee notably featuring two prominent female engineers. U.S. Navy Adm. Grace Hopper led the master committee to make a high-level language, while committee member Jean Sammet, who died May 20, 2017, led a subcommittee for the language itself.

Sammett worked for Sylvania at that time. She was responsible for the COBOL compiler of MOBIDIC (Mobile Digital Computer) being made on contract to the US Army Evans Signal Laboratory in Wall, NJ. Mobile was a relative term, as the computer required two 30-foot tractor trailers. This may have been the earliest computer to be programmed in COBOL.

Sylvania built six units. The computers were considered successful by Army users stateside and in Europe. System administrators, writing in 1973 of their experiences with MOBIDIC in the 1950s-1960s, said they were “pleasantly amazed by… the sophistication of the applications for which it has been used,” and that they would’ve kept using it if its hardware had not become obsolete.

SEE: Cobol: Grace Hopper’s gift to the world of business (ZDNet)

Today’s COBOL developers widely agree that the language keeps outlasting the hardware. IBM COBOL is of course the most popular version for mainframes. Another company, Micro Focus, is a leader in COBOL for distributed systems. There are other lesser-used flavors such as the open-source GnuCOBOL.

In any version it could take dozens of lines of COBOL for every one line of Java, but the latter just isn’t as fast for crunching troves of so-called big data, experts assert. Just about the only thing quicker would be moving data around a processor by hand. “If you really want to be efficient, go make sure you use the exact right assembler instruction code,” observed IBMer Maria Boonie, vice president for z Systems software development.

Still, “Starting about 2007 we went back and looked at this–is this rhetoric of COBOL being dead possibly true–and concluded by the numbers it’s simply infeasible. It’s not going away, it’s thriving,” Stoodley said.

Thriving may be a stretch, although Stoodley added that IBM expects the industry’s COBOL standards committee to regroup following its disbandment after a minor update in 2014. (TechRepublic’s messages to the committee convener went unreturned.)

Most customers write COBOL to maintain legacy applications. Stoodley said he does know of some new applications written in the language. Customers on both fronts require constant improvement from IBM’s compiler, which is currently in version 6.1.

Boonie noted that recent improvements focused on XML and JSON support, along with a continuous update model for the compiler itself rather than traditional software versions. What do customers still want? “We are starting to get some [interest] about 64-bit,” Stoodley said. “We’re starting to think about how it would manifest itself… there are some customers starting to run into that address space limit,” especially when parsing XML documents, Stoodley said.

Micro Focus for its part has a compiler in version 2.3 and plans to release 3.0 in July 2017, officials said. The company’s recent emphasis has been on modernization tools such as the Eclipse integrated development environment and Microsoft Visual Basic links. The summer release will have new features for debugging and additional platform support, they said. Micro Focus is also releasing a developer’s book online on June 6, 2017.

SEE: 50 years and still going strong: Will we ever be ready to kill off COBOL? (ZDNet)

COBOL programmer Mark Levy used the language at Sony and now works for United Parcel Service in northern New Jersey. There were already rumors of COBOL’s demise when his career started in 1984. Thirty-three years later, longer than most Java developers’ age, and a decade beyond the age of Java itself, he’s still at it.

“As a language it’s very easy to go in and make changes to existing legacy systems. Overall it really has been a steady language all these years,” Levy said. “You might have a system that’s done for a user on the internet and that’s written in Java or something else, but they end up calling back to COBOL accessing big mainframes and DB2,” he said, referring to IBM’s database management system.

Levy said he sees a decent amount of COBOL developers in their 40s and occasionally someone in the 30s. Many are born overseas, he added. IBM’s Stoodley said his company hopes to push that number younger with its Millennial Mainframer blog.

Disclosure: The US Army Evans Signal Laboratory today houses the InfoAge Science History Learning Center, home to various historical organizations including the nonprofit Vintage Computer Federation. VCF is led by the author.