The Donantonio system will work as the following rules:
Uses of Bibliographic Descriptions (metadata); each published resource needs an only document that describes it. For this, standard bibliographical schemes will be used (existing, or slightly modified or made).
The description does not have why to be hosted in the same server with the resource.
Each server will maintain his own descriptions, offering them to whom want to copy them.
It will exist formal differentiates between an original description and its copy. This difference will be shaped in an original-or-not mark. The originality mark only will have the original description, located in the publication server. Whenever an original description is copied the originality mark must disappear.
The fact to publish original descriptions or copies it's done in a combined way and without distinction.
The replicated descriptions (of different sources) can be put under a filtrate process to eliminate identical duplicates. It is to define the algorithm that distinguishes, for example, different versions, if it is necessary (is a filter that must have available only the end user or the query mechanism). Perhaps most ideal it is to have "modules of adaptable, interchangeable and modifiable filters". NOTE: This paragraph cannot be thus by anything of the world. .
A donantonio server could be also a donantonio client.
Each client will reply the descriptions of the servers whom he please, without maintaining no topology or concrete norm outside the free will of the administrator of the donantonio service .
At the moment it's not decided which must be the method of internal storage. Probably it is not necessary to never do it.
It's obligatory to import/export descriptions in the format defined by the Donantonio norm.
Query system between applications.
Techniques of end user to recover information:
search of patterns in description;
navigation by dynamic taxonomies;
bibliographical ontology (maybe you know what is that).
Processes for reply/filter/query (TO DEFINE).