You make a good point. In the evolution of systems, living or
non-living, it doesn't make sense to kill off the progenitors. Clayton
Christensen talks about "creative destruction", saying "occasionally
disruptive technologies emerge — innovations that result in worse
product performance, at least in the near term.....generally disruptive
technologies underperform established products in mainstream markets.
But they have other features....they are typically cheaper, simpler,
smaller, and frequently, more convenient to use." (02)
Ultimately, it is the the marketplace that decides. So we need to
understand how these conflicting feature sets interact with the
marketplace. In my philosophy courses I learned that "...you can't argue
a myth away. An existing myth can only be replaced with a better myth". (03)
I believe we are seeing the establishment of a new set of myths centered
around the cloud, big data and "intelligent search". Each has their
defining domain and each has their weaknesses in terms of ability to
meet expectations. (04)
I believe there is a time when the existing myths got us to a point for
new develop and some existing baggage must be jettisoned. We don't still
use gills but have replaced them with lungs that gave us access to more
oxygen than gills could. This allowed us to hunt prey in a metabolic battle. (05)
Likewise, in the computer and information arena we don't use the
techniques or software that was available many years ago. In the '70's I
was the manager of a computer store, one of the first, and the
predominant software was DOS based. Every day or so I would get a
customer coming to the store asking about software for their business.
They had friends who had commissioned software tailored to their
business. I was promoting the latest software, an automated spreadsheet
called "VisiCalc". I must have demonstrated it a hundred times. Still,
these prospective clients left the store and found some recent graduate
to code their application in a 1-off style. The new software that
eventually appeared did not migrate those early applications. They
disappeared as experiments in market economics. (06)
Likewise for biology. Most of the species that ever existed, are now
gone. Only the few most successful survive and they don't resemble their
What does this lead us to think about legacy systems? Well, we can say
that the current functions of legacy software will endure, they are
there for a purpose and that purpose will not go away soon. But many
applications will disappear and will be replaced by systems that can
address the same functions but using methods that the markets prefer. It
might be that intelligent systems will focus on problem solving and,
more or less, ignore how the software works. After all, Windows did not
migrate those early DOS applications but did note what businesses needed
for their applications. (08)
I believe the same is true of ontology based applications. I eschew the
extensive discussion of current tools because I know how ephemeral tools
can be. I'd rather spend my time understanding the required processing
functions because I know the useful ones will endure. We should look at
the requirements of legacy systems and focus on how ontologies can be
useful in those systems. (09)
Concord, MA (010)
On 8/30/2012 9:48 AM, John F Sowa wrote:
> The wonderful new opportunities for ontology, Big Data, and
> the Semantic Web are constant themes in Ontolog Forum. But I
> keep emphasizing the point that the new systems must accommodate
> the huge amount of legacy software that will not go away.
> I learned that lesson when I was at IBM. Following is a recent
> article about IBM mainframes (a few excerpts below).
> For another ancient legacy, just consider the Intel X86 architecture,
> which they considered obsolete in the early 1980s. It evolved from
> Intel's first microprocessor, the 4004. That had a 4-bit data path
> with a maximum RAM of 1K bytes. It was extended to the 8008, the
> 8080, the 8086, and the 8088.
> In the 1980s, Intel designed a clean new 32-bit architecture to
> replace it. But customers wanted upward compatibility, and Intel's
> shiny new chip was a failure. They continued with the 286, 386,
> 486, and Pentium.
> In the 1990s, Intel considered the Pentium the end of the road,
> and they developed the new Itanium, which was not a bad design.
> But it was incompatible with x86. So the latest and greatest
> new chips from Intel still have the oldest and ugliest detritus
> from the 4004 buried in their structure.
> Fundamental principle: Any revolutionary new designs for ontology,
> logic, Big Data, the Semantic Web, or anything else *must* support
> a smooth growth path from the old systems. The legacy systems can
> evolve to support the new, but they will never go away completely.
> John Sowa
> Excerpts from "I.B.M. Mainframe Evolves to Serve the Digital World"
> The death of the mainframe has been predicted many times over the years.
> But it has prevailed because it has been overhauled time and again. In
> the early 1990s, the personal computer revolution took off and I.B.M.,
> wedded to its big-iron computers, was in deep trouble. To make the
> mainframe more competitive, its insides were retooled, using low-cost
> microprocessors as the computing engine.
> Like any threatened species that survives, the mainframe evolved. It has
> been tweaked to master new programming languages, like Java, and new
> software operating systems, like Linux.
> “The mainframe is the most flexible technology platform in computing,”
> said Rodney C. Adkins, I.B.M.’s senior vice president for systems and
> The sale of mainframe computers accounts for only about 4 percent of
> I.B.M.’s revenue these days. Yet the mainframe is a vital asset to
> I.B.M. because of all the business that flows from it. When all the
> mainframe-related software, services and storage are included, mainframe
> technology delivers about 25 percent of I.B.M.’s revenue and more than
> 40 percent of its profits, estimates A. M. Sacconaghi, an analyst at
> Sanford C. Bernstein...
> A mainframe costs more than $1 million, and higher-performance models
> with peripheral equipment often cost $10 million or more. Yet even young
> companies and emerging nations, analysts say, find the expense worth it
> for some tasks.
> Comepay, for instance, is a fast-growing company that says it operates
> more than 10,000 self-service payment kiosks in Russia, where consumers
> pay for products and services ranging from Internet service and
> cellphones to electric bills. Comepay handles millions of transactions a
> day, and the volume is rising. The Russian company bought an I.B.M.
> mainframe in 2010.
> “Mainframes are extremely reliable,” said Ruslan Stepanenko, chief
> information officer of Comepay. “It keeps working even when the
> transaction load is very high.”
> Last year, the Senegal Ministry of Finance bought two I.B.M. mainframes
> to help monitor all the imports, exports and customs duties at the
> African country’s 30 border checkpoints...
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (012)