[Top] [All Lists]

Re: [ontolog-forum] Oooh, FOL is too hard to learn.

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Mon, 18 Oct 2010 17:52:58 -0400
Message-id: <4CBCC1BA.80605@xxxxxxxxxxx>
Ian, Rich, Pavithra, Adrian, Chris M, Ed, Chris P, and Ali,    (01)

Common Logic has no preferred concrete notation of any kind.
Any notation you prefer is, for you, a first-class CL notation.    (02)

>> ... people said  "Oooh, FOL is too hard to learn."
> Which is true. CL (pick any dialect you like) is difficult to work with.    (03)

That's false.  You have been using UML diagrams, and I assume that
you find them useful for your work.  When we discussed this issue
before, I was unaware that they had been defined as a CL dialect:    (04)

    http://www.omg.org/spec/FUML/1.0/Beta3/PDF/    (05)

> Ian, I agree completely.    (06)

> Difficult and easy is a figure of speech, and has no meaning
> if you do not have anything to compare it with..    (07)

> For lightweight natural language knowledge input and use,
> you may like to consider Executable English, as in...    (08)

If you don't like any of the dialects in the annexes of ISO 23707,
pick your favorite declarative notation, whatever it may be.  That
notation is, for you and your loved ones, a candidate notation for
Common Logic.  The next step to make that candidate a first-class
CL dialect is to specify a mapping to the CL abstract syntax.    (09)

> Maybe the CL/FOL community need to employ some HCI folks to help them
> develop the next generation of modelling notation, because what's there
> in CL now just isn't going to cut it with the average developer or data
> modeller.    (010)

> But now I'm confused.  Your claim is that it is first-order logic per se
> that is the problem.  If that is true, then it would make no sense for
> the CL/FOL to work with HCI folks to develop another notation, since it
> would be yet another FOL notation -- and hence yet another CL dialect
> and, hence, a notation that (according to you) no average developer or
> data modeler is ever going to work in.    (011)

> First-order logic is therefore only for the top 20% of data modelers,
> not the average...    (012)

I suspect that Ian and Ed are mistakenly equating FOL with a concrete
notation like predicate calculus or CLIF.  I keep telling people that
they've been speaking FOL all their lives.  Every sentence that uses
the words 'and', 'or', 'not', 'if', 'some', or 'every' is using full
FOL or some subset of it.    (013)

> I suspect, like Ian, that at least one contributing factor is that most the
> commercial programming languages have had quite a lot of effort devoted to
> making them easy to use to produce code. Whereas the effort in (notation
> for) logic has not been in the same direction.    (014)

A huge amount of effort has been devoted to UML, and programmers
have been happy to use it.  Given the fUML spec's, UML is CL.
The SQL WHERE-clause has the expressive power of full FOL,
and many programmers prefer SQL to XPath or SPARQL, even for
network-style databases.    (015)

Adrian has proposed EE, I've proposed CLCE, and Norbert Fuchs
has proposed ACE as English-like notations for logic.  Any or all
of them could be used as first-class notations for Common Logic.    (016)

We invite, encourage, and solicit anybody and everybody with
any HCI experience, training, or insights to design good
declarative notations.  Whatever they define and map to CL
semantics is a first-class dialect of Common Logic.    (017)

> I also don't see how FOL is inherently trickier than learning any
> given procedural programming language. Both require a certain
> perspective and background for one to become fluent in expressing
> themselves in. Admittedly, we have whole schools and courses
> dedicated from as early as high school to teaching people about
> procedural programming...    (018)

Dijkstra once said that the first prerequisite for a good programmer
is the ability to express oneself clearly one's own native language.
Huge numbers of people never learned that in high school.  But those
who can write clearly already have a head start in learning to write
FOL in any notation of any kind.    (019)

John    (020)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (021)

<Prev in Thread] Current Thread [Next in Thread>