Xref: utzoo comp.protocols.tcp-ip:7584 comp.dcom.lans:3099
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!husc6!m2c!jjmhome!cpoint!martillo
From: martillo@cpoint.UUCP (Joachim Carlo Santos Martillo)
Newsgroups: comp.protocols.tcp-ip,comp.dcom.lans
Subject: Re: Token Ring Patent
Summary: Ecrasez l'Islam et le patente de Soderblum
Message-ID: <2414@cpoint.UUCP>
Date: 3 Jul 89 18:30:14 GMT
References: <2377@cpoint.UUCP> <2413@cpoint.UUCP>
Reply-To: martillo@cpoint.UUCP (Joachim Carlo Santos Martillo)
Organization: Clearpoint Research Corp., Hopkinton Mass.
Lines: 166

In article <2413@cpoint.UUCP> martillo@cpoint.UUCP (Joachim Carlo Santos Martillo) writes:
>In article <2377@cpoint.UUCP> martillo@cpoint.UUCP (Joachim Carlo Santos Martillo) writes:
>
>I saw the following article a few weeks ago in comp.protocols.tcp-ip.
>
>>          Unfortunately, I know quite a bit about the Soderblom
>>          patent. It was first issued in the fall of 1981 after much
>>          consideration at the Patent Office. The senior patent in
>>          this area is Farmer and Newhall (AT&T). During the '70's,
>>          there were several interferences between the claims of
>>          Soderblom and those of Farmer and Newhall.  The Farmer and
>>          Newhall design, patented in 1971, was used by Dave Farber to
>>          help in the creation of the UC Irvine Ring (If I'm wrong,
>>          Dave, Please correct me). Then, Jerry Saltzer, Dave Clark,
>>          Ken Pogren and a bunch of people here at Proteon used the UC
>>          Irvine Ring to help in the creation of ProNET-10.
>
>>          Proteon never expected patent problems. In fact, in 1982, I
>>          presented details of our design at IEEE Headquarters in NYC
>>          to many AT&T employees including Bob Lucky. We joked about
>>          the fact that I was using BSTJ figures for my presentation,
>>          thus bringing new meaning to the phrase "not invented here".
>>          The Soderblom patent applied to systems which consisted of a
>>          master and a plurality of terminals. Farmer and Newhall was
>>          the senior patent when it came to closed loop operations.
>
>>          In 1983 or maybe 84, Soderblom presented at IEEE802.5. We
>>          were well aware of the Soderblom position. IEEE staff were
>>          also involved. We went ahead on the basis that Soderblom
>>          would provide licenses in a nondiscriminatory way. We also
>>          had a good idea of the probable terms for such a license.
>
>>          Since that time, the Patent Office has allowed a reissue of
>>          the Soderblom patent. The reissue has much more general
>>          claims. Many organizations filed objections which were
>>          brushed aside by the patent Office. That's why he has
>>          general coverage of a 1967 idea until 1998. His patents have
>>          run out elsewhere in the world. Don't blame Olaf.
>
>>          Many of us object to the whole procedure. That is, Was the
>>          token ring a new concept or an" old combination" in 1967?
>>          Many of us object to the details of the legal rangling.
>>          Nonetheless many of us have signed up simply because the
>>          patent office has spoken. In all cases, the exact terms of
>>          the license are covered by a non disclosure clause.
>
>>          It is also clear that some have chosen not to sign a
>>          license with Soberblom. It is expected that someone will
>>          take this to court for resolution.  Time will tell.
>
>>          Howard Salwen, Chairman-Founder
>>          Proteon, Inc.
>>          hs@relay.proteon.com
>
>
>I looked up the patent abstract in the Patent Office Official Gazette
>of October 6, 1981.  The engineering library at MIT has a
>subscription. On page 420 is found patent number 4,293,948, Data
>Transmission System, whose abstract follows.
>
>1.  Apparatus for the transmission of data characters in pulse form
>from a plurality of terminal units to a master unit comprising in 
>combination:
>
>	pulse input and pulse output means for each said terminal
>	unit and for said master unit,
>
>	a single series loop connecting said terminal units in
>	series along said loop between the pulse output means
>	and pulse input means of said master unit and over which
>	pulses originating either with said master unit or any
>	terminal unit are transmitted always in the same
>	predetermined direction,
>
>	each said terminal unit including:
>
>	(a) pulse responsive means connected to said pulse input means
>	for receiving pulses appearing at said pulse input means,
>
>	(b) decoding means distinctively responsive to pulses
>	received by said pulse responsive means,
>
>	(c) and logic means responsive to said decoding means for
>	normally interconnecting said pulse input and pulse output
>	means to thereby complete the loop at said terminal, unit,
>
>	(d) said logic means in response to a distinctive pattern
>	of said pulses appearing at the associated pulse input
>	means interrupting the loop between said pulse input and
>	pulse output means of stored data of arbitrary variable 
>	length as required by said terminal unit for transmission
>	over the loop to said master unit,
>
>	(e) said logic means being effective upon the opening of 
>	said at the first upstream terminal unit having data to
>	send for inhibiting the receipt at any downstream terminal
>	unit of pulses otherwise effective to control said downstream
>	terminal unit to transmit data.
>
>---------------------------------------------------------------------
>
>As usual, the language is totally obscure, and if the actual patent
>is similarly written, I would feel that the patent officer would have
>been negligent not to have rejected the application on those grounds
>alone.  I wrote away for the complete patent (it only costs a $1.50)
>but if this abstract contains the key elements of the patent, the
>issue of whether the token ring was a new concept or an old
>combination in 1967 seems totally irrelevant.
>
>From the abstract, Soderblum seems to have patented a circuit which
>is an intelligent unidirectional digital repeater/transceiver.  The
>intelligence involves the ability to become a transmitter of queued
>data upon reception of signals (a packet) on the receive input.
>Clearly, the 802.5 MAU is such an intelligent unidirectional digital
>repeater/transceiver.  Likewise, the FDDI MAU contains two
>intelligent unidirectional digital repeater/transceivers.  However, I
>doubt either MAU has any circuits borrowed from the Soderblom
>circuit.
>
>Now, I would consider the following scenario analogous to what
>Soderblom has achieved at the patent office.
>
>It is 1890 and I build the first electric stove which I patent.
>After I start selling my electric stove, other people begin selling
>electric stoves.  However, their heating elements, control elements
>and insulation are completely different.  In fact, other
>manufacturers begin manufacturing toasters and perculators and other
>devices which cook or prepare food using heat generated by
>electricity.
>
>I then go off and accuse all of these other manufacturers of patent
>infringement.  I would hope that the court would find my claims
>gratuitous except in the case of a stove manufacturer who had copied
>my original electric stove down to the nuts and bolts.
>
>In other words, I would have rights to my own design and that's all.
>Purporting to owning the concept of cooking with
>electrically-generated heat would seem a bit presumptuous.
>
>The Soderblum patent is equally presumptuous.  Unidirectional digital
>repeaters have been around since the 50s.  Adding a circuit so that a
>digital repeater receiving a signal from the receive input could
>become a transmitter is a nice little modification, and it might even
>have been slightly difficult using 1967 hardware logic and it may
>even be reasonable to give Soderblum a patent for that specific 1967
>circuit which no one in their right mind would ever use today.
>However, the trend to make relative dumb repeating, amplifying and
>filtering devices ever more intelligent through the addition of
>interpretive or program-type logic has existed in electronics since
>the first electronics engineers began working at the job and
>certainly has increased with the availability of microcontroller
>chips, PLAs and PLDs. 
>
>There is no obvious justification in granting patent rights over all
>circuits which constitute a class distinguished from a pre-existing
>class of circuits simply through minor extensions to enable some
>interpretation of incoming data with the result that the behavior of
>the circuits might be altered on the basis of interpretation.
>
>Joachim Martillo
>
>
>
>	
>
>