Github user Leemoonsoo commented on the issue:

    https://github.com/apache/zeppelin/pull/1265
  
    How about 
    
    
![image](https://cloud.githubusercontent.com/assets/1540981/18233684/0a788fb6-7331-11e6-88c7-1f50a023f78e.png)
    
    1) When "globally" is selected, "shared" mode is preselected and can not be 
changed
    2) When either "per note" or "per user" is selected, "scoped" or "isolated" 
is shown in the next stop down. And "+" button at the end.
    3) When "+" button is clicked, more drop boxes are shown to do combination 
of per note and per user.
    
    what do you guys think?
    
    
    Regarding per user interpreter config, i'm not sure how useful it is 
compare to complexity it will bring.
    
    Complexities per user interpreter config will bring are,
    
     - Some interpreter in 'scoped' mode might not able to support per user 
interpreter config. For example, current spark interpreter in 'scoped' mode 
shares single SparkContext. So there's no way to apply per user config which 
related to SparkContext creation.
     - Interpreter authorization will be more complicated to take care 
individual configuration's permission.
     - Interpreter configuration itself becomes complicated. For example, 
there're per user interpreter config created. When a user (devops, admin) want 
to change a property of all users's interpreter config at once, it's difficult 
to do that.
    
    I think the same goal of per user interpreter config can be achieved by 
improving interpreter authorization (permission), without much complexity.
    
    If interpreter authorization can control 
    
    a. who can create/clone new interpreter config
    b. who can edit interpreter config
    c. who can see and use interpreter config
    
    Then it can provide the exactly same thing that per user interpreter config 
trying to do.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

Reply via email to