Annotation of OpenXM/doc/ascm2001/design-outline.tex, Revision 1.5
1.5 ! takayama 1: % $OpenXM: OpenXM/doc/ascm2001/design-outline.tex,v 1.4 2001/03/08 00:49:29 takayama Exp $
1.1 noro 2:
1.2 takayama 3: \section{Design Outline and OpenXM Request for Comments (OpenXM-RFC)}
1.1 noro 4:
5: As Schefstr\"om clarified in \cite{schefstrom},
1.2 takayama 6: integration of tools and software has three dimensions:
1.1 noro 7: data, control, and user interface.
8:
9: Data integration concerns with the exchange of data between different
1.2 takayama 10: software or same software.
1.1 noro 11: OpenMath \cite{OpenMath} and MP (Multi Protocol) \cite{GKW} are,
12: for example, general purpose mathematical data protocols.
13: They provide standard ways to express mathematical objects.
14: For example,
15: \begin{verbatim}
16: <OMOBJ> <OMI> 123 </OMI> </OMOBJ>
17: \end{verbatim}
18: means the (OpenMath) integer $123$ in OpenMath/XML expression.
19:
20: Control integration concerns with the establishment and management of
21: inter-software communications.
22: Control involves, for example, a way to ask computations to other processes
23: and a method to interrupt computations on servers from a client.
24: RPC, HTTP, MPI, PVM are regarded as a general purpose control protocols or
25: infrastructures.
26: MCP (Mathematical Communication Protocol)
27: by Wang \cite{iamc} is such a protocol for mathematics.
28:
29: Although data and control are orthogonal to each other,
30: real world requires both.
31: NetSolve \cite{netsolve}, OpenMath$+$MCP, MP$+$MCP \cite{iamc},
32: and MathLink \cite{mathlink} provide both data and control integration.
33: Each integration method has their own features determined by their
34: own design goals.
35: OpenXM (Open message eXchange protocol for Mathematics)
36: is a project aiming to integrate data, control and user interfaces
37: with design goals motivated by the followings.
38: \begin{enumerate}
1.5 ! takayama 39: \item We should test the proposed standards mentioned above on
! 40: various mathematical software systems, but the testing has not been
! 41: enough.
1.1 noro 42: \item Noro has been involved in the development of
43: a computer algebra system Risa/Asir \cite{asir}.
44: An interface for interactive distributed computations was introduced
45: to Risa/Asir
46: %% version 950831 released
47: in 1995.
48: The model of computation was RPC (remote procedure call).
49: A robust interruption protocol was provided
50: by two communication channels
51: like the File Transfer Protocol (ftp).
52: As an application of this protocol,
53: a parallel speed-up was achieved for a Gr\"obner basis computation
54: to determine all odd order replicable functions
55: (Noro and McKay \cite{noro-mckay}).
56: However, the protocol was local in Asir and we thought that we should
57: design an open protocol.
58: \item Takayama has developed
59: a special purpose system Kan/sm1 \cite{kan},
60: which is a Gr\"obner engine for the ring of differential operators $D$.
61: In order to implement algorithms in $D$-modules due to Oaku
62: (see, e.g., \cite{sst-book}),
63: factorizations and primary ideal decompositions are necessary.
64: Kan/sm1 does not have an implementation for these and called
65: Risa/Asir as a UNIX external program.
66: This approach was not satisfactory.
67: Especially, we could not write a clean interface code between these
68: two systems.
69: We thought that it is necessary to provide a data and control protocol
70: for Risa/Asir to work as a server of factorization and primary ideal
71: decomposition.
72: \item We have been profited from increasing number
1.2 takayama 73: of mathematical software.
1.1 noro 74: These are usually ``expert'' systems in one area of mathematics
75: such as ideals, groups, numbers, polytopes, and so on.
76: They have their own interfaces and data formats,
77: which are fine for intensive users of these systems.
78: However, a unified system will be more convenient
79: for users who want to explore a new area of mathematics with these
1.2 takayama 80: software or users who need these systems only occasionally.
1.1 noro 81:
82: \item We believe that an open integrated system is a future of mathematical
1.2 takayama 83: software.
1.1 noro 84: However, it might be just a dream without realizability.
85: We want to build a prototype of such an open system by using
1.2 takayama 86: existing standards, technologies and several mathematical software.
1.1 noro 87: We want to see how far we can go with this approach.
88: \end{enumerate}
89:
90: Motivated with these, we started the OpenXM project with the following
1.2 takayama 91: fundamental architecture, which is currently described in
1.4 takayama 92: OpenXM-RFC 100 proposed standard %% ``draft standard'' and ``standard''
1.2 takayama 93: ``Design and Implementation of OpenXM client-server model and common
1.5 ! takayama 94: mathematical object format'' \cite{ox-rfc-100}.
1.1 noro 95: \begin{enumerate}
96: \item Communication is an exchange of messages. The messages are classified into
97: three types:
98: DATA, COMMAND, and SPECIAL.
99: They are called OX (OpenXM) messages.
100: Among the three types,
101: {\it OX data messages} wrap mathematical data.
102: We use standards of mathematical data formats such as OpenMath and MP
103: as well as our own data format {\it CMO}
1.2 takayama 104: ({\it Common Mathematical Object format}),
1.5 ! takayama 105: which can be expressed in terms of XML.
1.1 noro 106: \item Servers, which provide services to other processes, are stack machines.
107: The stack machine is called the
108: {\it OX stack machine}.
1.5 ! takayama 109: Existing mathematical software systems are wrapped with this stack machine.
1.1 noro 110: Minimal requirements for a target software wrapped with the OX stack machine
111: are as follows:
112: \begin{enumerate}
113: \item The target must have a serialized interface such as a character based
114: interface.
115: \item An output of the target must be understandable for computer programs;
1.2 takayama 116: it should follow a grammar that can be parsed with other software.
1.1 noro 117: \end{enumerate}
118: \item Any server may have a hybrid interface;
119: it may accept and execute not only stack machine commands,
120: but also its original command sequences.
121: For example,
122: if we send the following string to the {\tt ox\_asir} server
1.5 ! takayama 123: (OpenXM server of Risa/Asir) \\
1.1 noro 124: \verb+ " fctr(x^100-y^100); " + \\
125: and call the stack machine command \\
126: \verb+ SM_executeStringByLocalParser + \\
127: then the server executes the asir command \\
128: \verb+ fctr(x^100-y^100); +
129: (factorize $x^{100}-y^{100}$ over ${\bf Q}$)
130: and pushes the result onto the stack.
131: \end{enumerate}
1.4 takayama 132: OpenXM package implements the OpenXM-RFC 100 \cite{ox-rfc-100}
133: and 101 \cite{ox-rfc-101} based on
1.3 takayama 134: the above fundamental architecture.
135: In this paper, we discuss mainly on systems implementing
136: OpenXM-RFC 100 and 101 on TCP/IP.
1.1 noro 137: For example, the following is a command sequence to ask $1+1$ from
1.3 takayama 138: the Asir client to the {\tt ox\_sm1} server through TCP/IP:
1.1 noro 139: \begin{verbatim}
140: P = sm1_start();
141: ox_push_cmo(P,1); ox_push_cmo(P,1);
142: ox_execute_string(P,"add"); ox_pop_cmo(P);
143: \end{verbatim}
1.5 ! takayama 144: Here, {\tt ox\_sm1} is an OpenXM server of Kan/sm1.
1.1 noro 145:
1.3 takayama 146: Our project of integrating mathematical software
147: systems is taking the ``RFC'' approach, which has been
148: used to develop internet protocols.
149: We think that ``RFC'' approach is an excellent way and
150: we hope that other groups, who are working on standard protocols,
1.4 takayama 151: take this ``RFC'' approach, too.
1.2 takayama 152:
1.1 noro 153: The OpenXM on MPI \cite{MPI} is currently running on Risa/Asir
154: as we will see in Section \ref{section:homog}.
1.4 takayama 155: We are now preparing the OpenXM-RFC 102 ``Mathematical communication
1.2 takayama 156: on MPI'' (draft protocol)
157: based on our experiments on MPI.
1.4 takayama 158:
159: In the rest of the paper, we abbreviate
160: OpenXM-RFC 100 and 101 to OpenXM if no confusion arises.
1.2 takayama 161:
162:
163:
1.1 noro 164:
165:
166:
167:
168:
169:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>