polling for ripe dbase software changes (fwd)

Wilfried Woeber (woeber at cc.univie.ac.at) on March 19:
>   Hi Cengiz, hi David!
> 
> >We would like to hear the opinion (positive or negative) of the RIPE
> >database working group on the following proposals *before* changes are
> >made to any of the RIPE database implementations. We might want to
> >discuss things at the RIPE database working group session at the next
> >RIPE meeting in Berlin.
> 
>   Well, before I can develop an opinion, I'd be interested in the
>   motivation and/or expected benefits for the proposed modifications.
> 
>   My first *feeling* (note, nothing more than a feeling!) is that things
>   that belong to an external representation and a user interface should
>   not be part of the data stored in some database. 
>   Another aspect might be multi-language support. 
>   Of coure, you can translate the syntactic sugar into syntactic salt :-)

Currently, the database software is incapable of inserting the sugar
words back correctly (when the object has an error). Also, on rpsl,
syntactic sugar is actually syntactic delimiters. For example:

	as-out: to AS1 
	        to AS2 
	        to AS3 announce AS4
if you omit the delimiters:
	as-out: AS1 AS2 AS3 AS4
which is ambigious.

>   
>   BTW, my feeling (again) with regard to the continuation stuff is that
>   it might make sense to improve things. As it stands right now, it is
>   not really simple to comprehend (for humans).

Not by machines either. I think the line continuation rule is context
sensitive, so it can not be parsed by yacc/lex in a straight forward
fashion since they parse context-free languages.

> 
>   Give me more input, then I can make up my mind.
>   Best regards,
>   Wilfried.


Cengiz

-- 
Cengiz Alaettinoglu           Information Sciences Institute
(310) 822-1511                University of Southern California
http://www.isi.edu/~cengiz