RE: Status of LDIF and Changelog?

[Paul Leach <paulle@microsoft.com>]

> -----Original Message-----
> From: Chris Harding [mailto:c.harding@opengroup.org]
> Sent: Tuesday, November 17, 1998 1:51 AM
> To: ietf-ldapext@netscape.com
> Subject: RE: Status of LDIF and Changelog?
> 
> 
> Hi -
> 
> At 15:23 16/11/98 -0800, Paul Leach wrote:
> 
>   . . .
> 
> >Also, because XML is being rapidly accepted for use in the 
> Web, it is likely
> >to be very widespread. I was at Barnes and Noble, and there 
> are already a
> >half-dozen books on it.
> >
> This isn't necessarily the best criterion for making 
> something a standard
> (look at all the books there used to be on SGML :-). If you 
> could point to
> half a dozen major implementations of LDAP that used an XML 
> file format for
> data interchange, that would be a good reason to standardize it. At
> present, you can probably point to twice that number of 
> implementations
> using LDIF, which says to me that LDIF is the right thing to 
> standardize now.

We aren't disagreeing much. In my previous message, I said that a it would
be good to standardize current practice, but that I'd like to see
improvements go on the XML path.

> 
> Having said which, I agree that XML is potentially a better 
> basis for an
> interchange mechanism - at least as regards textual data (I'm 
> not so sure
> how it would cope with non-textual values).

It has provisions for non-text values.

 It has nicer ways 
> of delimiting
> entries and handling attributes, and it can cope with 
> non-ASCII character
> encodings. It would be particularly strong for textual 
> information if the
> major word processors were able to work well in XML-native 
> mode. This is a
> hurdle that they couldn't quite jump for SGML, but perhaps XML is
> sufficiently simpler that they can do it.

Like MS Office 2000? Its native document formats are XML.

> 
> As regards transformation of data from existing 'phone books 
> etc., XML is
> at present much less easy to work with than the LDIF format. This is a
> matter of the right tools being developed (I understand that 
> there are some
> PERL-based ones for SGML which probably work on XML also) and 
> of people
> gaining familiarity with them.
> 
> So perhaps there should be discussion on an XML format, 
> leading to Internet
> drafts and trial implementations, with a standard in prospect 
> in a couple
> of years' time?

I don't think it should take that long. If we were inventing a new
representation, I'd agree. But XML exists, and almost by definition, the XML
representation has to totally preserve the semantics of the ASN.1
representation. That reduces the task to one one of describing the rules for
making said transformations: non-trivial, but clearly easier than both
inventing a new representation _and_ creating such transformation rules.

Paul