You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Marshall Schor (JIRA)" <de...@uima.apache.org> on 2017/02/15 00:09:41 UTC

[jira] [Resolved] (UIMA-5162) uv3 change id to match v2

     [ https://issues.apache.org/jira/browse/UIMA-5162?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Marshall Schor resolved UIMA-5162.
----------------------------------
    Resolution: Won't Fix

keeping ids as incrementing, for now

> uv3 change id to match v2
> -------------------------
>
>                 Key: UIMA-5162
>                 URL: https://issues.apache.org/jira/browse/UIMA-5162
>             Project: UIMA
>          Issue Type: Improvement
>            Reporter: Marshall Schor
>            Assignee: Marshall Schor
>            Priority: Minor
>
> v3's current ids start at 1 and increment by 1.  v2 ids start at 1 and increment by the "size" on the main int heap of the FS. 
> Pros: v3's serializations / deserializations often need to recreate the v2 ids so the serialized forms can be read by v2 impl (for backwards compatibility), so this can have a performance benefit. It also improves the match for xmi/json id-refs with version 2 (although xmi deserialization should still work without this).  
> Cons: Changing to the v2 style will reduce the maximum number of Feature Structures per CAS somewhat; it will be the same as the current limit that exists in v2.
> Because of pros/cons, keep both impls, together with a global static final boolean switch to pick one or the other (and set it for now to use the v2 style). 



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)