Thanks very much for the response!

   - side-stepping tinkerpop was something I was considering for 
   establishing a class structure within the database and for facilitating 
   direct querying of the database by less technical users -- my first two 
   questions were aimed at allowing some on-staff analytics folks to query the 
   data repository directly without programmer intervention, without 
   developing a limiting layer of abstraction, and without having to teach 
   these folks gremlin
   - I want to avoid POJOs because I'm concerned about the long term impact 
   on code/data integration should the POJO structures need to change (which 
   is why a simple map-type property graph is appealing).. subsequenty, I'm 
   thinking there's not much cause for frames, true?
   - It's good to know that I don't absolutely need Rexster.  I'd be 
   tempted to consider it if I thought swapping out DBs was realistic (once we 
   go to production), but I don't.

Cheers,
Jon

On Monday, February 10, 2014 2:48:20 PM UTC-6, Jonathan Cook wrote:
>
> I am by no means an expert ... in fact still very much a novice, but I 
> think I can answer some of your questions.
>
> 1.  I believe you can still go lower level than Tinkerpop, but the 
> question is whether you'd need or want to for most use cases.
> 2.  I am definitely sure that you can use either or both.  (Plus "Pipes" 
> which is one level lower in the stack than Gremlin)
> 3.  I would think that you want to use the console or the OrientDB APIs as 
> I don't think that Tinkerpop exposes as much as you might want to leverage. 
>  I myself am creating my initial db with ODB SQL.
> 4.  No, Blueprints and OrientDB OO will play well together.  If you're 
> going to use Frames, check out TypedGraph.
> 5.  My database is not embedded and I am not using Rexster ... the only 
> thing obligatory if you want Tinkerpop is Blueprints, because that's the 
> layer where the individual Graph implementations are adapted to a common 
> API.
>
> Hope that helps and anyone with more or better insights (more experience) 
> please correct or expand :)
>
> Regards,
> Jonathan 'J5' Cook
>
> P.S.  I am looking to hire a Java/JS dev in the Washington DC area if 
> anyone is looking or knows someone!
>
> On Monday, February 10, 2014 10:47:30 AM UTC-5, Jon Machen wrote:
>>
>> Greetings!
>>
>> Thanks in advance for your patience with what may be an array of stupid 
>> questions.  I'm re-engineering a legacy enterprise application from the 
>> ground up and have down-selected to OrientDB for the underlying graph 
>> database.  After perusing this group, the various Wikis, etc I am left more 
>> than a little confused about where to get started.  More specifically:
>>
>>    1. Am I correct in my understanding that with 1.6.4 (or later) I will 
>>    need to leverage TinkerPop Blueprints in order to build a graph database?
>>    2. If so, am I obliged to use Gremlin or is the extended SQL provided 
>>    by OrientDB still available to me?
>>    3. Should I, ideally, be creating the initial (non-embedded) database 
>>    programatically through the TinkerPop APIs or establishing the structure 
>>    via OrientDB's native console?
>>       1. For whatever it's worth, a property graph IS precisely what I'm 
>>       looking to create.
>>    4. Does a blueprints implementation preclude me from using clusters 
>>    or classes (can I still use the inheritance structure built into 
>> OrientDB)?
>>    5. Given that I won't be using an embedded database, is Rexster 
>>    obligatory for a TinkerPop-based implementation?
>>
>> The majority of my experience to date has been with MS SQL and Neo4J (as 
>> is likely implied by my questions).
>>
>> Has anyone heard wind of a Rexster kibble to provide Cypher support? :D
>>
>> Cheers,
>>
>> Jon
>>
>>
>>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to