lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Doug Cutting (JIRA)" <j...@apache.org>
Subject [jira] Commented: (LUCENE-1473) Implement Externalizable in main top level searcher classes
Date Wed, 03 Dec 2008 19:57:44 GMT

    [ https://issues.apache.org/jira/browse/LUCENE-1473?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12652932#action_12652932
] 

Doug Cutting commented on LUCENE-1473:
--------------------------------------

> Doesn't Hadoop handle versioning inside of Writeable classes?

Currently, yes, but this probably insufficient for a Hadoop 1.0 release.  Hadoop is a distributed
system, and would like to provide RPC back-compatibility across minor versions after we go
1.0.  This is an explicit decision that's in the process of discussion within the Hadoop project.
 RPC's within Hadoop daemons will probably require identical versions -- all daemons in a
cluster must be upgraded in lockstep.  We'll thus probably limit back-compatibility to client
RPC's, so that a given client can talk to multiple clusters that are not running identical
versions of Hadoop.  Lucene has made no such explicit policy decision.  Lucene is not an inherently
distributed system.

Hadoop has not yet decided what mechanism to use to support back-compatible RPC, but Writable
versioning is not sufficient, since it does not handle RPC protocols, and it's lower-level
than we'd prefer.  We'll probably go with something more like Thrift or ProtocolBuffers. 
Hadoop does not use Java serialization and makes no promises about that.

> The developer contributions seem to be quite low right now, especially compared to neighbor
projects such as Hadoop.

As one who monitors both projects, I don't see a marked difference.  In both there are sometimes
patches that unfortunately languish, because, while they're important to the contributor,
they fail to sufficiently engage a committer.  For example, HADOOP-3422 took over 6 months
to get committed, probably because not many committers use Ganglia.

There is a difference in quantity: hadoop-dev has over 4x the traffic of lucene-dev.  But,
normalized for that, the number of patches from non-committers feels comparable.  If anything
I'd guess Lucene commits more patches from non-committers than does Hadoop.


> Implement Externalizable in main top level searcher classes
> -----------------------------------------------------------
>
>                 Key: LUCENE-1473
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1473
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>    Affects Versions: 2.4
>            Reporter: Jason Rutherglen
>            Priority: Minor
>         Attachments: LUCENE-1473.patch
>
>
> To maintain serialization compatibility between Lucene versions, major classes can implement
Externalizable.  This will make Serialization faster due to no reflection required and maintain
backwards compatibility.  

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org


Mime
View raw message