Descriptor store and User defined properties

[Dave Bryson <da...@miceda-data.com>]

Hi everyone,

I'm in the process of trying to layout a asset management system using
Slide.  One of the areas I'm having difficulty understanding is how I can add 
domain specific user defined properties for a resource other than the default
name-value props already in Slide. 

For example,  let say you have a book, chapter, and image and you want to
capture certain specific information about them.  I don't want to just chunk
that information into the current Slide properties table but rather put it
into a DB ( or search engine index ) that models a book, chapter, etc...  for
more efficient searching.

How best could you add this functionality to the current Slide system?

Is this were the future index service comes into play? In other words, you could
implement an index service that maps to these tables ( book, chapter, image )
and is controlled by the Slide api?  If so, where do you all see integrating
the index service helper - from the namespace token - like the other
helpers - or callable from outside the namespace token?   

Anyone doing anything similar to this with user defined properties that could
share some knowledge?

As a side note, my goal here is to use Slide for the basis of a digital asset
management system to manage potential large binary objects ( Word docs, PDF,
images, etc... ) that will allow users to add domain specific user defined
properties.  Of course, this will be an open-source application if anyone else
is interested. 

Thanks, 
-- 
Dave Bryson
daveb@miceda-data.com
----------------------

Re: Descriptor store and User defined properties

[Remy Maucherat <re...@apache.org>]

Hi Dave,

> I'm in the process of trying to layout a asset management system using
> Slide.  One of the areas I'm having difficulty understanding is how I can
add
> domain specific user defined properties for a resource other than the
default
> name-value props already in Slide.

In Slide the properties are namespace-name-value to allow easy translation
to / from XML. It wouldn't work well for binary content (if that's what you
want to do).

> For example,  let say you have a book, chapter, and image and you want to
> capture certain specific information about them.  I don't want to just
chunk
> that information into the current Slide properties table but rather put it
> into a DB ( or search engine index ) that models a book, chapter, etc...
for
> more efficient searching.
>
> How best could you add this functionality to the current Slide system?

I'll try to design and add a filter (TC 3 users would call it interceptor)
system soon. Feel free to contribute ideas on this if you're interested.

> Is this were the future index service comes into play?

I have the feeling that the answer is yes, and it would do what you want.

> In other words, you could
> implement an index service that maps to these tables ( book, chapter,
image )
> and is controlled by the Slide api?  If so, where do you all see
integrating
> the index service helper - from the namespace token - like the other
> helpers - or callable from outside the namespace token?

It will be like the other helpers. It doesn't add any restrictions because :
- You can directly acces the undelying indexer if you want to (for exemple
Lucene)
- Retrieving a namespace token is easy, and doesn't have a huge overhead
(it's just a facade object)

I plan to officially support two implementations :
- One which won't do anything
- One using Lucene (I looked at the API and some samples, and I liked it)

> Anyone doing anything similar to this with user defined properties that
could
> share some knowledge?
>
> As a side note, my goal here is to use Slide for the basis of a digital
asset
> management system to manage potential large binary objects ( Word docs,
PDF,
> images, etc... ) that will allow users to add domain specific user defined
> properties.  Of course, this will be an open-source application if anyone
else
> is interested.

Remy