Hi guys,

I just finalized TinkerGraph. The problem is that I don't want to be stuck 
having to keep TinkerGraph "as is" if people extend its classes. With 
GraphMutator coming in 3.1+, TinkerGraph internals will change. Finally, Titan 
was relying on TinkerGraph's Memory and MapReduce engine -- that would freeze 
TinkerGraph to always use those method, code. I have since copied the code over 
to Titan so Titan has a clone of the TinkerGraph Memory/MapReduce engine.

In TinkerPop2, we have had problems with people extending TinkerGraph --- its 
just hand cuffs I don't want to be stuck with.

Hope that is cool,
Marko.

http://markorodriguez.com

On May 20, 2015, at 2:10 AM, Bryn Cooke <[email protected]> wrote:

> This problem reminds me of the checkstyle rule:
> http://checkstyle.sourceforge.net/config_design.html#DesignForExtension
> 
> Could it be possible to determine which classes are likely to be subclassed 
> and add explicit extension points?
> 
> Bryn
> 
> 
> 
> On 20/05/15 06:56, pieter-gmail wrote:
>> I am less of a fan for using final.
>> 
>> For some future client it often turns a trivial solution using inheritance 
>> into a big discussion.
>> Sometimes the final is removed and other times a big engineering feat is 
>> required to keep it and solve the clients problem.
>> 
>> 2 cents,
>> Pieter
>> 
>> On 19/05/2015 23:59, Marko A. Rodriguez (JIRA) wrote:
>>> Marko A. Rodriguez created TINKERPOP3-692:
>>> ---------------------------------------------
>>> 
>>>              Summary: Should we final TinkerGraph?
>>>                  Key: TINKERPOP3-692
>>>                  URL: https://issues.apache.org/jira/browse/TINKERPOP3-692
>>>              Project: TinkerPop 3
>>>           Issue Type: Improvement
>>>             Reporter: Marko A. Rodriguez
>>>              Fix For: 3.0.0.GA
>>> 
>>> 
>>> If we are going to make TinkerGraph's classes final, we should do it before 
>>> GA. I'm all for making them final.
>>> 
>>> 
>>> 
>>> -- 
>>> This message was sent by Atlassian JIRA
>>> (v6.3.4#6332)
>> 
> 

Reply via email to