Sunday, February 20, 2005
BCNGroup Evaluation of
CoreSystem
On the Nature of science and
cultural Polemics à
On the discussion of data
regularity à
On the Stephenson Cyber Attack
Taxonomy à
(Download a 10 meg zipped
Macromedia presentation on CoreSystem.) à
communication from John
Sandy
You said:
The big contribution of CoreSystem is the manner in which it represents and manages Net-centric systems dealing with billions of nodes. Was that even a consideration in 1980?
In 1980, Arpanet had been around for over a decade, and IBM's VNET spanned the globe. In that year, I worked at IBM's Tokyo lab for 2 weeks, and I could log on to my VM/370 account at Yorktown (via a 3270 terminal) and run anything and everything I could do locally in Yorktown or at home via the telephone (including graphics -- 3270-style, which means character based, but with decent performance).
In the mid 1980s, IBM had supported the Andrew program at Carnegie-Mellon, which connected all the PCs and Unix workstations on campus (and via dialup or Arpanet) to servers that maintained all files. Anyone could walk up to any machine that supported Andrew, log on, and get all their files downloaded and uploaded as they needed them. Those who used it loved it.
Also in the mid 1980s, there were several proposals to extend a Unix-like file system globally. It would essentially be an Andrew-like system for the world. They were rather far along in their design, but the major deterrent was compatibility with the installed user base of mixed-vendor systems, including mainframes, PCs, and many different flavors of Unix, VMS, etc.
You said:
Our goal is to consolidate the computing paradigm and take the mainstream to a new level beyond the disjoint Java, UML, and XML models.
Certainly a good goal. But the major deterrent are two systems that were obsolete even by 1980 standards: Windows and X86. In 1980, Intel had already recognized that the 8086 had reached the end of its life and had a replacement, which died because of some idiots at IBM Boca Raton who chose the 8088 for their PC design (despite the fact that IBM already had a workstation in 1980 based on the Motorola 68000). Windows XP today still does not have the ease of use of the Mac of 1984 or the OS/2 of 1994.
I've seen many wonderful, elegant systems over the years, some of which I designed myself. But the biggest hurdle of all is getting anything new accepted that would disrupt the installed user base.
Fundamental principle: Nothing new will be accepted unless it is the path of least resistance.
John