System iNetwork

July 2006

« Previous Month Main Next Month »

July 6, 2006

Traversing a Paradigm Shift

It recently occurred to me that Thomas Kuhn's theory of intellectual history, where long periods of paradigm consistency are punctuated by shorter periods of 'paradigm shift', can be applied to the history of business computing. When I look at my own IT community -- the tradition of IBM computers progressing from the 360/20 and the Sys/3 to the iSeries and the system i -- I feel that we are still in the midst of traversing a paradigm shift, and the new period of stability has not yet settled firmly into place.

As I look back on my own career in IT (which began on a Sys/3 Mod 10), I see three eras of application design. The first era was the pre-workstation batch processing era of the 360 and the Sys/3. This was before green screens. Data usually entered the system through punched card readers, and data processing departments generally had more keypunchers than programmers. RPG II was the dominant language on the Sys/3, but there were no workstation files and little or no structured programming constructs. Nevertheless, we wrote effective applications which brought the power of computing to small and medium-size businesses.

Sometime in the 1970s there was a paradigm shift as a new era dawned with the introduction of the workstation. The Era of the Green Screen displaced the Era of the Punched Card. This new era ushered in the paradigm of the menu-driven application and interactive programs. This was new, this was cool, and I loved it! During the period of paradigm shift the old coexisted with the new: much data still entered the system through cards or diskettes because keypunching was faster than interactive entry of batch data.

Another characteristic of this second era was the triumph of structured modular programming over spaghetti code or the madcap innovations of creative programmer-artists ("Programming is an art," opponents of structured programming were heard to say). The addition of structured programming constructs to RPG II 1/2 and RPG III helped bring structured programming to the System/34-36 and the AS/400, but the limitations of the subroutine constrained the adoption of profound modular programming.

This second era stabilized and lasted for many years -- all through the 1980s and most of the 1990s. Indeed, it is still with us today, as the paradigm shift to the new era is not yet complete. We are in a twilight zone where two eras overlap, still traversing the paradigm shift.

The new era -- already here but not yet settled -- began in the mid-1990s with two radical new developments. First, IBM released RPG IV and brought us procedures and the ILE environment. This meant profoundly modular applications could now be composed in RPG. Second, the internet and the browser came on the scene. The browser is displacing the green screen as the dominant UI. The green screen is terminal and obsolesence is inevitable.

The Era of the Green Screen is passing away and the Era of the Browser is ascendant. Likewise, subroutine-constrained modularity is being displaced by procedure-based modularity. A new paradigm of application design is supplanting an old paradigm. This is simply what is happening, and resistance is futile.

A period of paradigm shift is a period of crisis. Uncertainty abounds, and "paradigm wars" occur between those attached to the former paradigm and those advocating a new paradigm. This is natural and it is constructive: the victorious paradigm will have been influenced and improved by the struggle and the critique.

For RPG, a specific crisis lies is the lack of a native GUI capability. Nevertheless, there are ways to solve this problem, such as CGI or a thin Java layer wrapped around an RPG core. And in a services-oriented architecture, our valuable RPG business logic can be preserved in modules which service the needs of an outer presentation layer.

Things will stabilize, and the new era may last as long as the preceding era did: more than two decades.

Posted by on July 6, 2006 at 10:14 PM | Comments (10)