printable multibyte encodings

[Joachim Carlo Santos Martillo Ajami <[e-mail]>]

   From: Theodore Ts'o <sgiblab!athena.mit.edu!troi![e-mail]>
   Subject: Re: printable multibyte encodings

      Date: Thu, 17 Dec 92 09:56:03 PST
      From: "Karl Auerbach, Empirical Tools and Technologies, 408/427-5280" <[e-mail]>

       > >Hmm.. Let's back up 20 years.

       > Let's back 8 up years. The development of the universal multi-octet
       > code just started within ISO/IEC JTC1 SC2/WG2. After many years of tremendous
       > efforts finally a multi-octet code was adopted by 80% of ISO voting
       > members as ISO 10 646 with an intention to replace the 7-bit ASCII -
       > ISO 646. 

      Hmmm... let's back up 500 years.  Folks decided litigation by seeing
      which side could come up with more supporters.  And countries decided
      issues with armies.

      The details may be different, but saying that a technology is good
      just because umpteen big, bureaucratic companies send umpteen "goers"
      to umpteen meetings where umpteen votes are held, bears the same
      degree of validity.

   Hear, hear.

   Furthermore, we can also back up 10 years, when people first started
   saying that OSI was "the networking technology of tomorrow", which
   everybody should drop and turn to right away, since its World Takeover
   was supposedly imminent.  :-)

						   - Ted

And I was pointing out 10 years ago that no one who had even the
vaguest acquaintance with the theory of political economics would ever
make such a silly comment that ISO OSI was "the networking technology
of tomorrow."  The ISO and its spiritual sibling the CCITT are
international regimes.  (A good reference is International Regimes
edited by Stephen D. Krasner from Cornell University Press).  As such
they never had any purpose to create or design "the networking
technology of tomorrow" (such a wonderful phrase).  International
regimes are tools for market manipulation and market protection
(sometimes also market creation and sometimes -- when someone's guard
is down -- market penetration).

In the ISO OSI case, most participants wanted to create a situation
where experience with providing networking solutions gave no country
comparative advantage in the world networking market and where an NTB
(non-tariff barrier) could be set up against foreign networking
manufacturers so that local networking industries could be fostered
which would then hold the local market against foreign competitors.

Such behavior is very typical in international trade.  Unfortunately,
in this case the Europeans managed to set up barriers which prevented
their own manufacturers from interacting with the US hi-tech
networking industry and from acquiring the expertise needed for the
successful development of such technology.  As a consequence, European
manufacturers have generally found themselves at a comparative
disadvantage relative to US manufacturers which can generally even
provide much higher quality ISO OSI technology than European
manufacturers.

In essence, the ISO OSI non-tariff barrier to the penetration of
European markets by US manufacturers has in a major turnabout and with
great irony become a major barrier to European penetration of European
networking markets.

ISO OSI will probably (with a few isolated exceptions) gradually fade
into oblivion over the next decade as it provides no benefit to any
but US ISO OSI providers and because ISO OSI technology is inferior to
several other forms of networking technology.

I would argue that using ISO OSI technology in the TCP/IP networking
arena simply because it is ISO OSI technology is probably a mistake.
In fact, SNMP's major flaw is probably the use of ISO 8824 ASN.1 and
ISO 8825 BER which like most of ISO OSI productions are needlessly
complex and do not partition problems well.  Certainly a much more
transparent network data definition language and a much simpler form
of binary encoding could have been developed.

In the current situation, I see a lot of relatively senior people
spending a lot of time defining network data objects (a task
appropriate to the first semester Harvard Extension School
introductory computer course).  I can only assume, they are spending
their time on this activity because their management is somewhat
baffled by ASN.1 and has not quite realized how trivial MIB definition
really is.

I am also astounded that comments like the following appear in
apparently respected networking texts and are apparently taken
seriously.

"The need for *non-proprietary* [my emphasis] networking technology
has long been recognized by several organizations, including the
*International Organization for Standardization/International
Electrotechnical Committee* (ISO/IEC).  In 1979, ISO in joint activity
with the *International Telephone and Telegraph Consultative
Committee* (CCITT), developed a reference model for Open Systems
Interconnection.

Anyone with the least expertise in internation trade theory and
practice would know that *non-proprietary* is the least of the
concerns of the participants in the ISO or the CCITT and that the
CCITT has been one of the major tools of internation market protection
for about a century.  I can only assume such comments are taken
seriously in engineering circles because of a lack of expertise in
fields of knowledge and learning beyond engineering.

Joachim Carlo Santos Martillo Ajami

The article represents noone's opinions other than the author's.