RE: Status of LDIF and Changelog?

[Chris Newman <Chris.Newman@INNOSOFT.COM>]

On Mon, 16 Nov 1998, Paul Leach wrote:
> > Paul Leach wrote: 
> >> I wouldn't mind having an LDIF standard, but IMHO it would be nice to 
> >> indicate that XML was the preferred way going forward. How many ways do
> >> we need to describe attribute/value pairs?
> > Well, since XML was introduced into the discussion, clearly N+1.
> 
> XML has pretensions to being a way to represent data in a way that is
> independent of application protocols. LDIF doesn't, AFAIK. In this respect
> XML is like ASN.1, although it is simpler, uses ASCII encodings, and defines
> (via XSL) ways to conveniently display itself. Or, you can think of it as
> S-expressions with angle brackets.
> 
> 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.

I'm baffled by this focus on syntax.  When it comes right down to it,
parsers for simple syntaxes like LDIF and XML-style-syntax are trivial
(often less than a day's work).

A few cases where syntax causes real problems are:
(1) if there is more than one syntax for the same thing
(2) if the syntax is hard to debug
(3) if the syntax has infrequently used options

LDIF has none of these problems.  Given that LDIF is already deployed, one
way to make things worse is to introduce an XML variant which does the
same thing.  That introduces problem (1) and results in interoperability
problems between vendors.  Possibly introduces problem (3) as well.

Therefore, I'm opposed to the notion of migrating from LDIF to XML at this
time.  It'd take some very substantive and specific engineering
justification to convince me migration is worth the pain.

		- Chris