Here in the US we're celebrating Labor Day. It's a Monday of a three-day weekend for a lot of laborers, although the day has turned into quite the commercial bonanza. It seems everyone wants to sell us mattresses and bedding sets this weekend. Perhaps sleeping season starts anew, with the end of the official summer vacation season.
While we ponder how much we owe to the historic labor organizations of the 20th Century -- things like national holidays, group benefits for health, the concepts of overtime and regulated reviews -- it's also a day to dig into the records for some 3000 history, too. I was tracking down technical papers for a 3000 consultant, one who'd asked the community to help him find his writing from the 1980s. I happened upon a paper from 25 years ago, offered at an Interex conference by HP's Orly Larson (at left). The genial advocate for databases was promoting the ideal of SQL for data storage and retrieval.
That might sound like advocating the benefits of sunshine or drinking water, but SQL was a long way from being essential to HP's 3000 success. It would take another five years, until 1993, for SQL to make its way into TurboIMAGE database architecture. In the meantime HP offered up three SQL products for 3000 DP managers. It was an era when the HP CISC processors, driving MPE V, were still in production use in the customer base. PA-RISC was laboring through its infancy among customer sites in 1988.
Larson sums up what was on the HP price list in 1988, and notes that Oracle was on the way for a late '88 release for MPE/XL, in a paper hosted on the OpenMPE website. The table (above) from that paper notes the first array of SQL solutions for HP's business computing customers. I've never encountered a 3000 customer who ever reported of using HP SQL. Allbase became a tick-box product for Hewlett-Packard while discussing 3000 options with new prospects. (Tick-box: yeah, we've got that. But nobody orders it.) Those customers who came in looking for SQL support on the 3000 were often convinced that the built-in IMAGE was a better choice, once you considered all the third-party software that was built to use that ubiquitous database.
There has been a lot of labor, across countless platforms, to elevate SQL selection to the equivalent of turning on a spigot for a drink of fresh water. Other technologies that seem new today, and have pending impact on MPE use like cloud computing and virtualization, will experience those years of laboring to become de-facto standards. The labor comes from the integration aspirations of IT managers, working overtime on long weekends like this one, to deploy something lauded but not fully proven. SQL was once a laborer in that state.