Re: UFN take 2

[Sylvain Langlois <[e-mail]>]

> One more justification to all my previous comments
>     . single keyword value  for each attribute
>     . mandatory "," delimiter (even at line break)
>     . recommended "bracket" delimiters

I missed the original posting from  Steve and the  first comments from
PAP (due to Fnet Gateway congestion I guess -- this  is the very first
OSI-DS message I  received since  last  week!).  would some  body mind
resend these two directly to me please? Thanks.

> Within our OPAX X.500 pilot project we have some (limited but
> the hard way) experience of "multi-vendor" X.500 implementations,
> and we really badly need a unique common way of representing
> attributes and attribute values AT EVERY LEVEL

>      . for human exchange of DN
>      . for documentation (technical or not)
>      . even for text based exchange of Directory objects


> In our last meeting, we even decided to try to work out
> a proposal for such a representation.

I'm not sure we really need to use this representation at every level.
I've tried to work a bit on this and came to the fact that the less we
have to "translate" the better we can do it! I recently posted  a mail
on this list but probably lost again for the same reason  as above!. I
think we can easily achieve a common representation of a subset but we
have to have the translator working in both ways (in and out). This is
rather simple for text-based representation  (QUIPU, Pizarro), this is
going to be rather diffucult for others (DIR/X). I have to look closer
for Custos and SQL based stuff which is going to be another problem!


Sylvain

----------------
Sylvain Langlois		  "Dogmatic attachement to the supposed merits
([e-mail])	   of a particular structure hinders the searc
h
				   of an appropriate structure" (Robert Fripp)