Re: What is an algorithm?
Date: Sun, 21 Jun 2015 14:10:40 -0700 (PDT)
Message-ID: <0201df62-488e-4554-8992-3647c28a9f1a_at_googlegroups.com>
> In this field about algorithms Y. Gurevich and N. Dershowitz, collaborate in
> their papers. However Dershowitz in his paper "Effectiveness" writes:
> "States must be comprehensive: they need to incorporate all the relevant
> data, when coupled with program..."
> Much earlier before this paper I have defined states as a general knowledge
> about entities or relationships. Note that the entities and relationships
> are much more general categories than algorithms and programs.
1.
Related to the above-mentioned Dershowitz generalized idea, I would like to
say that it is done entirely in my papers. I introduced the states in my
paper from 2005. I developed completely the theory of states. Related to
states, I introduced a theory of identification and the general theory of
databases. In my paper from 2008, I introduced the concepts of states as
well as a number of other things. I would like to underline that I
introduced the states of entities and relationships. The entities and
relationships are most general concepts.
2.
In this paper, "Effectiveness", Nachum Dershowitz proposes that
"Constructors" should be introduced, look at section 3.2:
"In general, then, the operations in states come in three flavors: domain
constructors, defined functions; and black-box oracles.". Later in this section, he introduces a definition of "Effective state":
Definition 1 (Effective State)
- A state is basic if it includes constructors for its domain, ...
(for more details about "Effective State", look at Derskowitz's paper "Effectiveness", section 3.2)
In my paper from 2005 ("Some ideas about a new data model", section 2.1), I
introduced "Constructor" and "ClosingConstructor." I use similar names as in
object oriented programming, but in fact, my procedures differ from OOP
procedures. First, notice that I do not have "Destructor" because there is
no deleting of data. Another thing, my constructor entries complete
knowledge about one data. When it comes to programs, then "general
knowledge" in my theory of states is really general knowledge. This means
that when it comes to the states of programs, which are placed in a
database, then there may be remembered for some relevant data from the
compilation, operating system etc. So there is indeed genaral knowledge.
Note that I have defined knowledge as a set of atomic facts, that is the
corresponding factual sentences.
Why Constructor and ClosingConstructor are important? There are many
reasons, I will mention two of them: With these procedures is completely
determined, who's working operations with data. The names of these
procedures we can put in the db so that we know for each data; which
procedure has entered that data in the db. We can notice that for the entry
of a data item we can apply different procedures. For example, dates can
operate three different Constructors.
The second reason for the application of these procedures can be a complete
automation at the level of data entry.
Vladimir Odrljin Received on Sun Jun 21 2015 - 23:10:40 CEST