You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@stanbol.apache.org by "Rupert Westenthaler (JIRA)" <ji...@apache.org> on 2011/06/27 11:12:50 UTC
[jira] [Created] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Create a "stable" launcher for Apache Stanbol
---------------------------------------------
Key: STANBOL-241
URL: https://issues.apache.org/jira/browse/STANBOL-241
Project: Stanbol
Issue Type: Improvement
Reporter: Rupert Westenthaler
Assignee: Fabian Christ
The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
This should also help to identify components to be included in a first Stanbol release
Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Issue Comment Edited] (STANBOL-241) Create a "stable"
launcher for Apache Stanbol
Posted by "Olivier Grisel (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055566#comment-13055566 ]
Olivier Grisel edited comment on STANBOL-241 at 6/27/11 2:11 PM:
-----------------------------------------------------------------
I don't want to discontinue the stateless launcher: it is very useful when you want fast, offline processing (without network latency to any kind of remote server).
Typically:
org.apache.stanbol.enhancer.engines.opennlp.ner
org.apache.stanbol.enhancer.engine.entitytagging
+ small preloaded data for dbpedia.
For the following engines:
org.apache.stanbol.enhancer.engines.metaxa
org.apache.stanbol.enhancer.engines.opencalais
org.apache.stanbol.enhancer.engines.zemanta
we need to align them all on the stanbol vocabulary which is not implemented yet.
was (Author: ogrisel):
I don't want to discontinue the stateless launcher: it is very useful when you want fast, offline processing (without network latency to any kind of remote server).
Typically:
org.apache.stanbol.enhancer.engines.opennlp.ner
org.apache.stanbol.enhancer.engine.entitytagging
+ small preloaded data for dbpedia.
For:
org.apache.stanbol.enhancer.engines.metaxa
org.apache.stanbol.enhancer.engines.opencalais
org.apache.stanbol.enhancer.engines.zemanta
We need to align them all on the stanbol vocabulary which is not implemented yet.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Issue Comment Edited] (STANBOL-241) Create a "stable"
launcher for Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055578#comment-13055578 ]
Fabian Christ edited comment on STANBOL-241 at 6/27/11 2:45 PM:
----------------------------------------------------------------
Okay - we can still have the stateless launcher and just add a stable launcher.
For stable I would use this set at the moment:
Stable:
- Enhancer
- EntityHub
Unstable:
- ContentHub + SPARQL Endpoint
- FactStore
- Ontoloymanager
was (Author: fchrist):
Okay - we can still have the stateless launcher and just add a stable launcher.
For stable I would use this set at the moment:
Stable:
- Enhancer
- EntityHub
Unstable:
- ContentHub + SPARQL Endpoint
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Rupert Westenthaler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057883#comment-13057883 ]
Rupert Westenthaler commented on STANBOL-241:
---------------------------------------------
regarding the definition of "stable":
For me "stable" is also a question of how stable the features provided by this set of components are.
about Components:
LangId is for me no candidate for the stable launcher, because it has a LGPL dependency that is very unlikely to be resolved. Therefore it will most likely be not in a stanbol release. We should not include functionality in the stable launcher that is than missing in a release.
Metaxa is quite useful and it would be a good to have it in the stable launcher. However in the long term it would need to be refactored from an Engine to the a Content Adapter. I think that even for a release it would be ok to keep Metaxa as an Engine, therefore this should be no argument for excluding it from the stable launcher.
Regarding OpenCalais, Zamanta and Geonames: This engines all need an API key *). We need simple to decide if we want to include those engines in the stable launcher or not. As long as they are "unsatisfied" without the required API key they do not really hurt and if users do have the required key it would be much easier for them to start using them.
In any case we need to validate that the Enhancements created by such engines are in conformance with the Enhancement Structure because we can not aspect from users to adapt there clients only because the are activating a different set of Engines.
E.g. I know that the Zemanta Engine can not provide entity-types values for EntityAnnotations something that may be a real issue for users. Geonames should be ok in that respect. I have not had a detailed look at the OpenCalais Engine.
best
Rupert
*) This is also true for geonames, because there is the possibility to create a free account and than use the api.geonames.org server that is much more stable than the one that allows anonymous connections.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055578#comment-13055578 ]
Fabian Christ commented on STANBOL-241:
---------------------------------------
Okay - we can still have the stateless launcher and just add a stable launcher.
For stable I would use this set at the moment:
Stable:
- Enhancer
- EntityHub
Unstable:
- ContentHub + SPARQL Endpoint
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Olivier Grisel (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057203#comment-13057203 ]
Olivier Grisel commented on STANBOL-241:
----------------------------------------
We also need to pre-fill the entity hub with a small solr index of the most important dbpedia entities. Rupert is working on one.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057819#comment-13057819 ]
Fabian Christ commented on STANBOL-241:
---------------------------------------
I would like to stay with the term "stable" launcher and come up with criteria that define when something qualifies as "stable". I agree that at the moment perhaps none of the available enhancement engines are really stable.
For LangId, Metaxa, and OpenCalais I haven't tried them or seen examples of their usage. So are they stable in the sense of not raising problems because there are no problems or they are not used? Perhaps one criteria for "stable" would be that there are examples and a GUI for their usage.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Rupert Westenthaler (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057219#comment-13057219 ]
Rupert Westenthaler commented on STANBOL-241:
---------------------------------------------
I am also moving the configurations for DBpedia (referenced site, solr index, entity tagging engine for dbpedia) from the launchers to the default data bundle.
This will have the advantage, that one needs only to update it in the default data and not in all the different launchers. I will link the commit with this issue
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Walter Kasper (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057825#comment-13057825 ]
Walter Kasper commented on STANBOL-241:
---------------------------------------
Just use the full launcher and query the metadata after sending a file for the results of the respective engines (for OpenCalais, of course, you first have to supply your API key). That had been demonstrated at Istanbul last year. Ther have been also documentation pages on the IKS-Fise Wiki with examples, now moved to Stanbol-README's. With OpenCalais, you might even switch off the openNLP engine and just use the Contenthub etc GUIs to see similar results, even for non plain text content.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Olivier Grisel (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055566#comment-13055566 ]
Olivier Grisel commented on STANBOL-241:
----------------------------------------
I don't want to discontinue the stateless launcher: it is very useful when you want fast, offline processing (without network latency to any kind of remote server).
Typically:
org.apache.stanbol.enhancer.engines.opennlp.ner
org.apache.stanbol.enhancer.engine.entitytagging
+ small preloaded data for dbpedia.
For:
org.apache.stanbol.enhancer.engines.metaxa
org.apache.stanbol.enhancer.engines.opencalais
org.apache.stanbol.enhancer.engines.zemanta
We need to align them all on the stanbol vocabulary which is not implemented yet.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Issue Comment Edited] (STANBOL-241) Create a "stable"
launcher for Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055556#comment-13055556 ]
Fabian Christ edited comment on STANBOL-241 at 6/27/11 2:46 PM:
----------------------------------------------------------------
I reviewed the available enhancment engines and would suggest to use the following distinction in stable / unstable
Stable:
org.apache.stanbol.enhancer.engines.opennlp.ner
org.apache.stanbol.enhancer.engines.geonames
org.apache.stanbol.enhancer.engine.entitytagging
Unknown:
org.apache.stanbol.enhancer.engines.metaxa
org.apache.stanbol.enhancer.engines.opencalais
org.apache.stanbol.enhancer.engines.zemanta
org.apache.stanbol.enhancer.benchmark
Unstable:
org.apache.stanbol.enhancer.engines.langid
- only one test for english. what about other languages?
org.apache.stanbol.enhancer.engines.autotagging
- should be replaced by an engine that uses EntityHub
Please comment especially on the engines with unknown status. What criteria do we have to call an engine stable?
was (Author: fchrist):
I reviewed the available enhancment engines and would suggest to use the following distinction in stable / unstable
Stable:
org.apache.stanbol.enhancer.engines.opennlp.ner
org.apache.stanbol.enhancer.engines.geonames
org.apache.stanbol.enhancer.engine.entitytagging
Unknown:
org.apache.stanbol.enhancer.engines.metaxa
org.apache.stanbol.enhancer.engines.opencalais
org.apache.stanbol.enhancer.engines.zemanta
Unstable:
org.apache.stanbol.enhancer.engines.langid
- only one test for english. what about other languages?
org.apache.stanbol.enhancer.engines.autotagging
- should be replaced by an engine that uses EntityHub
Please comment especially on the engines with unknown status. What criteria do we have to call an engine stable?
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Fabian Christ resolved STANBOL-241.
-----------------------------------
Resolution: Fixed
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Walter Kasper (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057804#comment-13057804 ]
Walter Kasper commented on STANBOL-241:
---------------------------------------
What is called "stable" here, probably should better be called a "minimal" launcher" as opposed to the "full" launcher. To us, the proposed "stability" classification apparently has little to do with the concept of software stability in the sense of not raising problems, complaints etc every other day. LangId, Metaxsa and OpenCalais are very stable in that respect, while I would regard "geonames" as being "unstable" (due to the networking problems, regularly demonstrated at demos). Also, entitytagging looks to us more "unstable".
So I suggest to change the name from "stable" launcher to "minimal launcher".
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13057824#comment-13057824 ]
Fabian Christ commented on STANBOL-241:
---------------------------------------
I have removed the geonames engine from the stable launcher and added the launcher to the top-level POM.
Will resolve this issue for now as the launcher now exists. The discussion about what is "stable" in Stanbol should be made on the list. Then we can add/remove components based on the agreed definitions.
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (STANBOL-241) Create a "stable" launcher for
Apache Stanbol
Posted by "Fabian Christ (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/STANBOL-241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055556#comment-13055556 ]
Fabian Christ commented on STANBOL-241:
---------------------------------------
I reviewed the available enhancment engines and would suggest to use the following distinction in stable / unstable
Stable:
org.apache.stanbol.enhancer.engines.opennlp.ner
org.apache.stanbol.enhancer.engines.geonames
org.apache.stanbol.enhancer.engine.entitytagging
Unknown:
org.apache.stanbol.enhancer.engines.metaxa
org.apache.stanbol.enhancer.engines.opencalais
org.apache.stanbol.enhancer.engines.zemanta
Unstable:
org.apache.stanbol.enhancer.engines.langid
- only one test for english. what about other languages?
org.apache.stanbol.enhancer.engines.autotagging
- should be replaced by an engine that uses EntityHub
Please comment especially on the engines with unknown status. What criteria do we have to call an engine stable?
> Create a "stable" launcher for Apache Stanbol
> ---------------------------------------------
>
> Key: STANBOL-241
> URL: https://issues.apache.org/jira/browse/STANBOL-241
> Project: Stanbol
> Issue Type: Improvement
> Reporter: Rupert Westenthaler
> Assignee: Fabian Christ
>
> The Idea is to discontinue the "stateless" launcher and convert it to a launcher that only include features that are stable.
> This should also help to identify components to be included in a first Stanbol release
> Use this issue to discuss what components to be included
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira