Re: CE Developing a Controlled English

[Adam Pease <apease@ks.teknowledge.com>]

David,
   An excellent and practical post.  I'd like to declare myself a 
participant.  Although there are other approaches or details I'm sure I'd 
agree with, your outline sounds fine to me in its entirety.  I'll suggest a 
starter grammar in a separate post.

Adam

At 08:32 PM 5/31/2001 -0500, David Whitten wrote:

>I would like to initiate some discussion on this mailing list.
>The stated purpose of this list is:
>Purpose:: To discuss and develop a Controlled English as the language of SUO.
>
>(this is from our home page: http://suo.ieee.org/ )
>
>While I understand that any language can be ambiguous in more than
>one way, I'd like us to limit ourselves to the following guidelines:
>
>  1) Discussions must be strongly related to developing SUO-CE.
>
>  2) Any thread, whether on-topic or off-topic may be tabled by
>     request of five or more participants.
>     A separate message requesting tabling a thread is discouraged.
>     A tabled thread may be resumed after being quiet for one week.
>     Any message about a tabled topic is counted as continuing the thread.
>
>  3) Working code or computable algorithms will be preferred to
>     general statements of intent.  Any programming system may be used.
>     Commonly available systems will be preferred.
>
>  4) a person may declare themselves a participant in the SUO-CE
>     working group explicitly, or implicitly by asking for a thread to
>     be tabled.  Simply posting messages to the list does not make
>     someone a participant.
>     Participation is by individual, not by organization.
>
>  5) Anyone may declare themselves to not be a participant in the SUO-CE
>     working group explicitly, or implicitly by cross-posting a message
>     to both the SUO-CE mailing list and any other SUO mailing list.
>     Politeness suggests that they refrain from posting any further messages
>     to the mailing list in that case.  All participants are encouraged
>     to ignore any postings by non-participants, until they declare
>     themselves as participants.
>
>  6) Since the purpose of the working group is to design and implement
>     the SUO-CE system, a regular report will be posted back to the
>     main SUO group outlining our progress.
>     No more than one report will be sent every two weeks.
>     A draft of that report will be sent to the SUO-CE mailing list
>     two calendar days before it will be sent to the main list.
>     If the thread discussing the draft is tabled, no report will be sent.
>     If the thread discussing the draft is active within two hours of
>     the time to send to the main list, no report will be sent.
>
>  7) A thread will be started to determine who is willing to act as
>     an officer of the mailing list meeting.
>     The officers include:
>
>Enumerator : responsible to note when a thread has been tabled.
>   Reporter : responsible to send the report of progress
>   Rosteror : responsible to answer private emails about participant roster.
>
>Of course, this organization of the mailing list is up for discussion.
>I'd hope that we can think of any other ideas and clarify this quickly.
>
>The rest of this e-mail is about the SUO-CE itself.
>
>How to make a SUO-CE:
>
>1) The grammar for SUO-CE should be a context free grammar.
>    If we find sample sentences that we want to parse that require
>    a context sensitive grammar, we will list them,
>    but not address them in the first phase.
>    the first phase should be parseable by some open-source Earley Parser.
>
>2) We will have two kinds of lexical items:
>    a) Defined in the Lexicon
>    b) Inferred from the input
>
>3) There will be a strict distinction in our discussions
>    between a word as an input to the controlled English parser: "word"
>    and the structure after parsing which is its meaning:  [word]
>
>4) The mapping in English between a "word" and a [word] appears to
>    be many to many.  The Controlled English mapping must be computable,
>    and predictable to most .
>
>5) The goal of SUO-CE is to end up with a system that provides a more
>    English-like syntax than the traditional Pierce-Peano notation for
>    formal logic.  SUO-CE will be a logical notation, however and not
>    attempt to handle all of the ambiguities of unrestricted English.
>
>
>What does this group think ?
>
>David

Adam Pease
Teknowledge
(650) 424-0500 x571