GoogleGazer

When I began work for IBM in 1970, mainframes were where it was at, “Complex Systems” (interconnected mainframes) was the ultimate state-of-the-art, the highest-speed character-based terminals were connected to mainframes at a speed of 9.6 kbs (about 1,000 characters a second on a good day), and PL/I an imperative computer programming language was the latest word in rapid development techniques.

Over time, Python, C#, C++, and Java replaced Assembler, PL/I, COBOL, and Fortran, speeds of processors and communications increased exponentially, “graphics” supplanted character-only terminals, memory and storage costs plummeted, and “distributed computing” became in vogue, putting more computing power closer to the user, and efforts focused on trying to get things off the “ancient” and hard-to-maintain mainframes. Then came the PC revolution, and for a while decentralization was all the range, and mainframes became very much “old school” and passé, their death warrants having been signed by columnists long ago. Then, people began to figure out two things: (1) they were putting a lot of complexity onto desktops, and this engendered mucho frustration, and required an expensive army of technicians and “help desks” to maintain some semblance of usability, and (2) those desktops had lots of power that went mostly unused most of the time. As Internet service became more reliable, as connection speeds increased, and as specialists like Google and Amazon learned to build multi-acre “farms” of cheap computers reliably and inexpensively, wise men and women figured out that “software as a service” (SAAS) had many potential benefits, and the transparent distribution of the centralized services was imperative to keep pace with wildly fluctuating changes in demand. Thus was born “cloud computing” which is the current rage.

The wizened old GoogleGazer is, once again, reminded, as the french proverb says, plus ca change, plus c’est la meme chose (the more things change, the more they remain the same). In computing, things stay the same, but the terminology changes, the mechanics of execution get slicker, and the costs of hardware decline inexorably as Moore’s law has has proved itself for a half a century now. But the bottom line seems to be that we have now come full circle, and for good and for ill, we are moving with alacrity back to the era of centralized software hosting, centralized applications, and centralized control.

This is not necessarily a bad thing. The fundamental needs addressed by the grandfathers of today’s young tiger developers haven’t changed too much. People still seek powerful but simple to use and reliable solutions, and for the most part, they just want to get their work done without worrying too much about the complexity under the covers. As Samuel Clemens (”Mark Twain”) famously observed as he got older, he was amazed at how much his dad had learned in the years between when he himself was 16 and 21! Not all old ideas are bad ideas, and implementing them with new technology can be a downright good idea.