directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Karasulu <akaras...@apache.org>
Subject Re: A little bit of organization
Date Wed, 06 Dec 2006 15:08:04 GMT
Emmanuel Lecharny wrote:
> Alex Karasulu a écrit :
> 
>>
>> Hey,
>>
>> Yeah I used to have it that way a few years ago as a matter of fact. 
>> However there was not much there then so I consolidated.
>>
>> Perhaps it's time to break things apart again.  However what's the 
>> motive for the package reorganization?
> 
> Well, I just started to write some more syntaxChecker (as we need some 
> of them for the metaSchema), and as we have something like 30 SC to 
> implement (as of RFC 4517), I just think they deserve a specific 
> package, because we will also have around 30 normalizers, and 30 
> comparators, so around 100 java files in the same package seems to me a 
> little bit too much.

Sounds like a good reason.  Same one that I thought but perhaps you had 
other reasons as well.  Let's do it.

> However, I also think we must discuss that a little bit further, because 
> we might want to allow ADS administrators to add their own syntax 
> checkers, which will induce some different packaging.

Yeah more than just admins.  We want people that can manage schema 
(allowed to based on ACI) to be able to do that which entails adding 
more of these code components: syntaxCheckers, comparators, and normalizers.

So they can use any package name they like.  Also the way we will 
represent them in the meta-schema and the subscheme entries will be 
interesting.  Namely we will keep their bytecode in the metaschema entry 
and represent these entries as attributes within the subschemaSubentry.

This will require some custom class loading which we need to discuss. 
First thing first, we need the standard syntax checkers, a meta-schema, 
and schema entity description parsers for these entities.

Glossary:

*metaschema* => schema to represent schema

*subschemaSubentry* => attribute that points to one massive entry 
containing a bunch of attributes that store schema entity descriptions 
in attributes

*schema entity description parsers* => basically we have attributes in 
the schema subentry like syntaxes, attributeTypes, objectClasses, 
matchingRules etc where each value of these attributes uses a special 
syntax to describe a schema entity.  In the syntaxes attribute a syntax 
description is provided.  In matchingRules a matchingRule description is 
provided.


HTH, let me know if there is more that needs clarification so other 
understand us.

Regards,
Alex



Mime
View raw message