Automated Reflection and Aggregated Persistance

[Xittenn]

I have been working on my computing engine for many moons now. I have chosen to adopt COM as a technology to support the infrastructure of my core API. As such I have also chosen to utilize Structured Storage as the method by which I will maintain object persistence.

 

Structured Storage presents much in terms of POD type Metadata and opens up the possibility, at least in my opinion, for an embedded or aggregated system by which Persistence can be driven by Reflection. I have received some flack for my thinking as Structured Storage and OLE isn't intended for use as an intensive Persistence technology. The main advantages I see to doing so is the long term support of the technologies and the already integrated nature of Structured Storage with COM.

 

So why use Automated Reflection to Aggregate Persistence in C++ a language not structured for Reflection? Well the first and foremost answers involve addressing issues with speed. Moving forward on this C#.NET and F# are more readily integrated on top of a core that is structured on COM and C++ thereby enabling the structuring of managed code and functional language driven by the underlying horsepower of C++. It also allows for intense graphical representation of simulations through the DirectX API. And above all I understand the language.

 

So why is any of this important for anything? What do I hope to accomplish through Automated Reflection and Aggregated Persistence? I hope to be able to put something together that enables Artificial Intelligence and better and more flexible simulations for my purposes. For much the same reasons CERN has adopted similar strategies to manage their data systems. Much in the same way as CERN has developed their system I hope to develop my own and despite its present lack luster state I hope that it will one day be readily capable of parsing data and integrating new systems and pathways to better understand the data it is being presented with.

 

I have been through a few forums and the conversation has been COM, Persistence, Reflection, not so much. My response is usually well I'm not listening to you. So I am presenting here the basics of why it is I feel I am doing right by my application in hopes to hear the underlying reasons why others feel this is wrong. I may add to my statements further when more comes to mind and I have intentionally remained vague and without examples as I hope to start something and move on it from there.

 

Thanks,

BekaD:

[Nex6]

I am not really deep enough to go into all the C++ and and the more formal models. but, I try and stay away from com,OLE etc. but then again, I am an engineer hacker....

 

in one high distributed software project we wrote own own helper agent to collect an xml file that the main process generated. the helper agent then could send it back to the webservice to be process into the DB. but, that's allot slower then you need. but for us was very efficient.

 

maybe its my dual, windows and Unix side but I like to keep things simple....