Versionable binary content best practices

[Robert Morse <rm...@mdesignz.com>]

Hello,
I'm looking for an example of saving documents (pdf files) in the  
repository along with various properties.
These documents need to be versioned.
I've found a couple of references, but nothing that really addresses  
the issue.   During my search I came across a post
on jcrdev.com that concisely asked what I'm trying to do (although the  
poster wasn't concerned about versioning and I am).

Any help would be greatly appreciated, and once I produce a solution,  
I'll be more than happy to post it so others can benefit.



<------ Post from jcrdev.com ------->

hello all,

Although it sounds like a basic question, I just can not find a good  
source to answer, what I think is such a basic problem.

*Produce binary content.

*have properties associated with that content.

*want to store, then be able to search the properties and retrieve the  
binary content.



------

Example:

Document1.pdf

client_fname: bart

client_lname: simpson

client_company: yoville

document_type: contract

document_sign_date: 5/5/2009

-----

use cases:

*Search for all documents associated with bart simpson, print them all  
out.

*Search for all contract documents for company yoville.

*Search for sign_dates older than 10 years ago, archive off.

What should the CND/nodetype/primary or mixin setup be to support the  
above, especially if you plan on handling 100k of those documents  
(narrow v deep structures)?  I've seen nt:unstructured/nt:resource/ 
nt:file (sometimes with recommended nt:folder above nt:file) and then  
they say 'add properties as desired' but no REAL examples.

TIA, just suprised I can't find examples






The human mind is an iterative processor, it never does anything right  
the first time.  What it does well is to make improvements on every  
iteration (deMarco)