Re: Re: software design clarification

Markus Elfring wrote:
> 
>> Let's take "abstraction" as a process, starting with concrete use cases,
>> and looking for commonalities. So one question becomes: "What programming
>> languages can be represented by the Synopsis AST ?".
>>   
> 
> I'm also interested in this one. - I've tried to suggest a top-down
> design approach while you are asking for a bottom-up design here. I hope
> that we can meet somewhere in the middle.

:-)

I'm not sure I propose a bottom-up design, actually. What I try to point
out is that Synopsis does have a design, and whatever is missing is driven
by the C++ specification. That's the context in which we should discuss.

You seem to keep arguing for a design that needs to be done, but you fail
to tell us what it actually is that you want to design. Think "use-case driven
design", if that helps.

> Can a common class library be achieved that will be implemented in some
> target programming languages?

common to what ? And why the future tense ? What about the existing APIs ?
Have you actually looked at them ? How do they relate to what you are asking
for ?

> When are you going to rename parts of your software from "Tree" to
> "Graph"? (I guess that the separation of concerns between "T" and "G" is
> another nasty source of confusion. When is the "S" related to the terms
> "syntax" or "semantic"?)

No idea. To be honest, such a name change is the least of my concerns right now.

>> AFAIU, swig does quite the opposite from Synopsis. It takes a high-level
>> interface description, and generates wrapper code.
> 
> Is C/C++ high-level enough for an IDL?   ;-)
> http://www.swig.org/Doc1.3/Introduction.html#Introduction_nn10

OK, fine. That paragraph suggests swig does some code analysis, too.
What was your question, then ?

>> IDL is nothing but a language. There are various language bindings that
>> define the mapping to language-specific interfaces. So this, too, is a
>> simple
>> matter of code generation, not code inspection.
>>   
> 
> How is the current state for the support of the CORBA interface
> definition language?
> http://synopsis.fresco.org/docs/Tutorial/idl-parser.html

It's almost as complete as the the omniidl IDL parser in the most
recent release (http://omniorb.sourceforge.net/). There is at present
no support for value-types, merely because I haven't looked into how to
represent them in the AST.

Feel free to contribute.

>> What 'libraries' do you have in mind ? The code implementing a C++
>> compiler
>> frontend surely isn't related to python at all.
>>   
> 
> Front-/Back-ends can be written in Python and C++.

I don't understand what you are saying, and how it relates to the rest
of the discussion.

Regards,
		Stefan

-- 

      ...ich hab' noch einen Koffer in Berlin...