http://www.noologie.de/symbol03.htm (URL)

The main focus of the discussion in the present text is a factor of information system design that finds very little attention in the current software industry: the crucial factor of user interaction speed. It is said that humanity as a collective is very bad at remembering lessons from the past. One saying goes: He who doesn't remember history is forced to repeat it forever [22]. In the just about 50 years of history of the computer industry, this applies even more than to any other sectors of society, and the forgetting time lag seems to be much shorter than anywhere else. In those 50 years, there were several waves or phases of computerization, the mainframe era, the mini era, and the micro era. It seems as if the lessons learned in one phase were hardly ever heeded in the next one, and the most serious errors were repeated. Very little information flow seems to have occurred between the people designing the systems of those different phases, and with the present-day WIMP technology, many advantageous and even essential factors of the prior generations were thrown overboard.

Back in the olden days of timesharing mainframe computers, the user often had to wait a long time for responses from the overloaded mainframe system. For a while, the personal computer brought a respite from this situation. But with the current generation of behemouth operating systems residing on every desktop computer, it looks like the situation is very much the same. The user input responses are unbearable if one doesn't have the most powerful, fastest, and of course, most expensive hardware in the desktop box [23].

The most important design goal of the computer industry is to make the systems useable by the largest number of people. The WIMP technology is designed exactly for this end. That would be allright, if only there were provisions that made the systems more usable for expert and power users. WIMP SW needs much higher computing resources than command line SW, but even if those resources are there, it still is about one order of magnitude slower than command line SW. This is due to the time penalty incurred by the necessary mouse fiddling. These factors are treated with in more depth in the following chapters.

What WIMP technology completely obliterates is a phenomenon called "neuronal feedback loops" [24]. The crucial element of "neuronal feedback loops" is the human short term memory. If most interactions take place within a time lag span of about 1/10 sec, there arises a "flow effect" (Csikszentmihalyi) that seems to lead to an expanded short term memory capacity. Since this is very hard to measure, and most scientific quantitative measuring setting often work against the grain of such an effect, this has found very little documentation in the technical literature, and even less application in commercial products. My own experience arises from many years of experimentation and design with systems development.

There seems to be a vast, and almost completely unknown, territory of augmentation of human potential lying dormant behind this screen of "neuronal feedback loops". There is a different dimension to the old adage that "information that doesn't arrive in time is no information". It applies as well to the human short term memory.

The design maxim in the present discussion is: make it as fast as possible! This is obviously an agenda that goes against the grain of the tendencies ruling the present computer industry. And that is the deeper reason why this work is being undertaken.

[22] The source of this is, I believe, from Santayana.

[23] This looks very much like a marketing trick performed by a coalition of SW and HW vendors (the WINTEL cartel) and could give rise to a host of conspiracy theories. My personal experiences with the MS Win 95 system are dismal. On a system with a Pentium 90 processor, Win 95 is an exercise in frustration. The task switch reponse times are in the order of one to ten seconds. More on this in the later sections.

[24] I am using this name of my own design. IBM fellow Tadhani had discovered this in the mainframe generation, and Csikszentmihalyi had reported on this, not in connection to computers, but as a general phenomenon, as "flow effect". Its scientific description is still wanting, and I have not found very much literature. Ernst Pöppel (neurology) has done published some essential work, but this has no connection to practical applications in the computer industry. The name of "neuronal feedback" stems from the observation that a stimulus-response-stimulus loop between a human operator and a system needs to be maintained within a time frame that is close to the flicker fusion time of a cinema presentation (the 1/25 sec time lag when single pictures fuse into an illusion of motion). Most symbolic interaction systems of the past never cared about such time factors, and therefore these aspects pertaining to higher symbolization functions and rational processing have found very little attention. Only with the advent of suitable computing systems can these effects be transferred to symbolization. Therefore the name "Symbolator".