A recap of a new user's journey into trying to learn OrientDB after a week 
at it.

*The good*

   - I see a lot of great features in OrientDB which I think are really 
   useful
   - I find the Udemy class and documentation do a reasonable job. There 
   are areas for improvement, but overall one can find most of the basics.
   - I see a lot of potential and areas where I think OrientDB would be a 
   good match. Not coming from a graph background there is much  in terms of 
   concept I am still figuring out, but it seems manageable.

*The bad*

   - 
   - I can only guess the sql layer was created to help people familiar 
   with SQL to use OrientDB. However, there are a number of inconsistencies 
   which make it confusing and ad to the learning curve. Anywhere were it was 
   easy to keep the same syntax/behaviour as sql it should have been kept. 
   What may appear as flexibility is, in my opinion, can contribute to feature 
   overload, make it harder to learn for new users and cause more work to 
   maintain for the development team.
   - Number of features, sometimes some of which in my opinion are 
   redundant, makes it harder to learn. Some of the work I see on 
   orientdb-labs seems to address some of this. For example the unified 
   Multi-model.
   - Storage is confusing. After years of using Postgresql recently started 
   to look at NoSQL engines. Spent a few months learning Cassandra and have it 
   in one small project to learn the ropes. I found their storage straight 
   forward to understand. You have replication levels, read/write consistency 
   and data is sharded. In the week I have been looking at OrientDB I have 
   found storage to be one of the most confusing parts and also an area that 
   needs work. From what I can tell data/indexes are not sharded yet. I 
   believe Issue #6256 is about that. Even the definition of cluster in the 
   manual is confusing... 
   in https://orientdb.com/docs/last/Tutorial-Clusters.html it mentions it is 
   loosely like a table, but from everything I have seen a class seems closer 
   to a table in the RDBMS world. I still can't quite figure out if a cluster 
   is the equivalent of a tablespace or if it is more like a "parent table" ( 
   table made up of inherited tables with the same structure so one can query 
   directly a single table or the parent table).
   - Online presence for OrientDB is not only confusing, but also 
   frustrating: Google group suggests to use StackOverflow right on top; so 
   what was this group created for then? I wrote an email to 
   [email protected] and got no response. I did not expect that to 
   go into a black hole. I asked on this group about training and got no 
   response. Given the, very low, volume of posts in this group I am hard 
   pressed to see why someone from OrientDB Ltd could not respond; at a 
   minimum I expected someone to respond to the email.  It is a very poor 
   image and makes one wonder if one went with an enterprise license if one 
   would get the same poor level of response. Right now I am looking at 
   OrientDB for an internal project, but I would be very hard pressed to 
   recommend it to any of my clients right now because I would be afraid we 
   could not get support if we needed it. Specially the clients that would not 
   have an issue with cost, would be the ones that would also be most 
   demanding about making sure we have support.

-- 

--- 
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/d/optout.

Reply via email to