If this were true, you should be able to find many Usenet postings that use this term to describe the software. We have the archives. The article indeed attempts this search and doesn't find it.
There is an unrelated older meaning of "open source" from "open source intelligence", but the idea that the meaning of "open source" as "source visible" was widespread before 1998 seems false. Perhaps some people had heard of it, but it was not widely understood.
Indeed, in the article, two of the Usenet hits for "open source" are "open source material", which can be disregarded as the older meaning of "open source intelligence", which means to collect data from people from things like phone books. The other hit is of dubious meaning.
There is certainly an event horizon wherein computing platforms became entirely too sophisticated to support 'open source' everything - at a very early period of computing history, perhaps - but I do believe that source code has always been open. Its the fact of code going closed, multiple times, and now persistently, that is confabulating everything. I maintain this is the inverse function.
I spent three days trying to get one of the very early versions of Mosaic to compile on an SGI Indy and never did get it to work. They didn't have a binary release at the time because getting that to work would be even harder.
Around .. '84 / '85, there used to be these great little systems from MIPS that ran Risc/OS. Any time we'd get an update to Risc/OS, we'd spend a few days rebuilding drivers and whatnot, for our production systems - these drivers were "open source", in that they required us to compile them ourselves. (I think they were from the PROGRESS RDBMS product...)
Back in the days of USENET, any code that was posted to groups like comp.lang.c was considered 'open source' - it was just a description of the sources, whether they were available or not. There wasn't any attribution of value - whether it was free or not - just whether or not we had access to it .. or not.
Nowadays there is all sorts of stigma and hubris around the subject - but back when computers were something you had to visit in a special room designed for the purpose, all it really meant was whether we had permission to read the source - and do things with it - or not. There wasn't a commercial value assignment, really, until the mid-90's, when people realised there was immense value in open source business models (RedHat, et al.) ...