On Monday 18 February 2008, Stuart Brorson wrote:
> Dare I suggest that some type of XML or XML-like tree
> structured format might be a good thing?  I don't necessarily
> like XML, but I do appreciate the advantages of having a file
> format which is extensible, i.e. we can add new tags while
> not breaking older parsers.  Maybe there's a way to do that
> using a line-oriented file format?

XML specifies only a syntax, which is not very important.  You 
could just as well say to make it C-like, Lisp-like, ADA-like, 
or whatever.  It doesn't matter.

What does matter is two factors ..

1. Organization of the information, the ability of the format to 
represent the data appropriately.

2. A precedent for its use, ideally for the exact use.  It is 
likely you won't find an exact precedent for its use, so the 
next choice is for related precedent for its use, where there 
is a logical mapping to this use.

By a "related precedent for its use", I mean to look at related 
tools like auto-placement, static timing analysis, signal 
integrity simulation, schematic, and others.  You might also 
consider import and export to competitive tools to be in this 
category.

Of these companion tools, some already exist, some will be done 
if we plant the seed.

You need to decide whether to use a portable format, that might 
allow direct sharing, or to make another proprietary format, so 
translation would always be required.

When a new format is chosen, the first coding that needs to be 
done is a two-way translator, between the new and old format, 
which will prove the adequacy (or inadequacy) of the format.  
It needs to be able to make a round trip with no loss.


_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev

Reply via email to