vendredi 24 juin 2011

Data mapping perspective

From all the beginning, since the first world summit on cybernetics in 1945, data structure and schemes has been a very controversy subject. Many approaches have been experimented, and the concept of a unique approach (or specification) is still fa away to reach.

From a researcher perspective, It's rather a usage issue than a structure issue, cause any specification will face the dilemma of interoperability. It's nearly impossible to come up with a versatile product or API. Such technologies have been a rich subject of research within giant corporations such Google (REST/ATOM), Microsoft (SOAP/XML), Sun (RPC/XML), and others.

Many think that XML have been chosen mainly for it's interoperability features. More than that, there are more important features to consider : Processing and Mapping.

XML was developed as a data specification to embrace higher goals ; emphasizing simplicity of creation and manipulation, language transcoding (trough Unicode and UTF-8). Not only, XML benefits from other specification that allowed easy validation and processing such as DTD, RelaxNG, XSLT and XPath. All that allowed XML to used in very complex data processing project (such as the Genome Project - ref: EMBL bank formats).

But XML is not the only specification to gain such interest. SQL, Javascript, and many others are the technologies that marked history of information technologies.

From a developer perspective, I've been facing more complex issues than that (as many of my colleagues) ; Writing application that can embrace multiple data formats, multiple protocols isn't something simply affordable to every developer, especially facing the growth of the web.

Many solutions are encountered through the web ; mainly as abstraction layers, such as :

  • Data source : ODBC, JDBC, PHP DB API, Python DB API ...
  • ORM : Doctrine, SqlAlchemy, Hibernate, ...

But, I don't think it's sufficient. That is the reason why a came up a the idea of building a specific software piece ; whose only task is to allow is to handle data retrieving, processing, validation, manipulation and presentation. It's codename ICE Hive, and is included in the ICE Framework. I wont tell that would solve all the problems, but I can assure that will make your life easier when it come to inter-source data computing.

Aucun commentaire:

Enregistrer un commentaire