To: Charles Hughes who wrote (21740 ) 11/25/1998 1:06:00 AM From: rudedog Read Replies (1) | Respond to of 24154
Chaz - Maybe I was getting a little cosmic behind some excellent cognac - but I think actually there are just some minor semantic differences between what you are saying and my thoughts. I agree with every one of your data points, but I think there is a trend implied by that data (with a few other odds and ends thrown in) which led me to the conclusion that browsers are a transitional step. Although the concept of a graphical user interface goes back at least to the PDP-9, and some pretty sophisticated ones were in use on CAT scanners more than 20 years ago, the general user interface to the innards of the computer world up until the mid-80's was via text - VT100, 3270, or some such thing. Graphics as a part of the computer environment was limited to a few very expensive special purpose products. I would generally credit the Xerox Star project as the first attempt at a mainstream GUI - and it incorporated many of the features we now take for granted (since both the MAC and Windows borrowed so heavily from the concepts). And exclusively graphics-oriented interfaces really did not become well accepted until the early 90's. Anything like real multi-media capability is more recent yet (some would say still in the future!!). So I would say that based on the percentage of users who regarded a GUI as their primary interface, the reign of the GUI is less than 5 years old. GUIs as a mainstream thing happened at almost exactly the same time as the emergence of the first versions of Mosaic from U of I. the browser is a kind of GUI I agree 100% - but in the sense that a GUI is a kind of user interface (like text terminals) and text terminals are really just another input device (like punch cards or paper tape). Each step up in abstraction introduces a shift in the sense of where a user is. When a user toggled in a boot program on the front panel of a 1440 or PDP-8, he knew exactly where he was and what he was interfacing with - registers which were about to hammer those toggled bits a million times faster than the human had just done. Punch cards, paper tapes, and the early file systems allowed one to assume a level of abstraction just above the machine. Text terminals, real-time multitasking and early networking took us higher yet - things began to be called by name and users interfaced with virtual things which were only loosely connected with their underlying resources. The mass market GUIs just built on that metaphor without really extending the reach of the user. But the browser and the web behind it extended the user to yet another level of abstraction, one in which locality played a much less prominent part. One of the most captivating things about the web is this sense of community and communication which extends beyond the boundaries of any local control, and the browser is the interface which brings this to the user. For this and other reasons I believe that browsers are about to eclipse a local GUI as the primary interface for most users. But there is a separate parallel shift going on, not in the interface but in the environment that users will interface with. Some visible pieces of this are storage area networks and the shift in storage concepts that will be driven by truly virtualized persistent data; the shift of policy domains from an essentially physical model to one based more on where services are provided; the shift from a grounded procedural model where execution occurs at a known place to one where the requestor and provider may both be hosted by resources which aside from execution of the transaction, have no particular connection to the origin or target of the request. You are right of course, the user experience with browsers, and the underlying meta-language, will not immediately change. Users and programmers are just beginning to understand what we have. But browsers arose naturally in parallel with the web, eclipsing the earlier text-based internet mechanisms without really replacing them, just as initially no one thought of browsers as a replacement for a local GUI – as you say, it is in fact an application which runs on that GUI. It's just that increasingly, the other uses for the local GUI are being marginalized. My premise is that although you are right, the browser metaphor has just begun to achieve mainstream acceptability, and huge momentum (and inertia) will be generated by that acceptance, the seeds of that shift were already sprouting at the point where the previous generation user interface was achieving that same mainstream status. The next shift will not take as long as the last one – in fact, each generation of interface technology seems to last about half as long as the previous one, a kind of inverse corollary to Moore's law. If that trend holds for a while longer, browsers had better hurry up and conquer the world – they've only got about a 2 year run from this point. I have a number of ideas on where ‘this stuff might pop up' and I am testing some of those assumptions now, but nothing really well enough baked to discuss on an open forum. I doubt that you will run out of interesting challenges any time soon, however, and as far as I can tell, you are in no danger of seeing any significant portion of this thinking appear in the University course list any time soon.