Re: Suggestions required please .... (MarkAGr)
10 Apr 1999 18:15:13 -0400

          From comp.compilers

Related articles
Suggestions required please .... (1999-04-03)
Re: Suggestions required please .... (1999-04-06)
Re: Suggestions required please .... (Dobes Vandermeer) (1999-04-06)
Re: Suggestions required please .... (1999-04-09)
Re: Suggestions required please .... (1999-04-09)
Re: Suggestions required please .... (1999-04-10)
| List of all articles for this month |

From: (MarkAGr)
Newsgroups: comp.compilers
Date: 10 Apr 1999 18:15:13 -0400
Organization: AOL,
References: 99-04-036
Keywords: design

This is the second post that informs me that my use of english is rather
obscure .... I'll attempt to clarify using "rspartan"'s post as a template ...

In article 99-04-036, writes:

>HI, found your post and find it interesting, but complicated.
>I'm working on a grammar tool similar to UN*X Lex that attempts to
>generates a program on any language (Pascal, C, Java). It seems
>similar to your problem.
>So, let me traslate your question to english. ;-)
>1. You have a grammar tool that reads a data description language

I am writing an object based system description tool ...

>2. Will generate a program or library on another language (just "any"
>language) like assembler, embedded, C, Java.

Yes, the purpose of the tool is to convert the semantics of a "system
description" into an alternative language ... usually a programming
language, often a diagram, occasionally english prose, very
occasionlly (but I suspect it will become more frequent) meta-data
languages. I've even had one enquiry to produce a description of gates
- for hardware synthesis.

>3. You NEED a intermediate language or intermediate code.

Yes ... an intermediate language to describe a procedural description..

The tool consists of a number of grammars .... those seen by the
user(system developer), those used by the developer(tool developer)

Grammars used by the user (system developer)

A system description is split into three parts, each are textual descriptions:
        a) An information model.
                a description of each object, its attributes, its linkage/navigation to
                other objects in the system
        b) A message interaction model
                a description of what messages are passed from object to object
                (messages may be either synchronous or asynchronous )
        c) A procedural definition model (PDL)
                description of what happens when a message is sent or recieved.

The above descriptions are created either :
        efficiently, in a text editor, by a user(system developer) or
        inefficiently by a graphical tool (that looks like UML, SM, SY,
        or our own graphical representation)

Grammars used by the developer (tool developer)

The tool internally consists of
        A) a parser of (a,b,c & fwd:C)
                that extracts the semantic content and stores it in a database(fwd:B)
        B) a database ( used to be an rdb but now ... written in (a) )
                to hold the semantics of (a,b,c & C)
        C) a generation script language
                an interpreted script to generate the final ( alternative language )
                output from (B)

I have (a) & (b) .... and am asking for help with a definition for (c)
I have (A for a), (A for b) & (A for C) ... I will need an (A for c) as soon as
c is firmed up...
I have (B) and am willing to radically modify it to reflect the contents of (c)
I have (C)

        I need a procedural definition language .... I have looked at ....C, C++,
SDLpr, PASCAL ... thanks to others in this forum I will be looking at MOF, ODL,
ASL ... (ASL seems the closest but I don't want to copy it. I want to conform
but have problems in the "re-use vs plagerism" debate. (I'm not sure that
copying this language 'straight' is the way to go, especially as many of the
concepts included for SM are redundant)) One option is to allow generated-code
augmentations to be defined in the language that is being generated, however
this does lead to a "non-portability" of the model.


OK .... help !!!!! suggestions please.
All these memories will be lost in time, like tears in rain ...

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.