uima-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jörn Kottmann (JIRA) <...@uima.apache.org>
Subject [jira] Commented: (UIMA-1875) ability to visualize and quickly update/add values to primitive features
Date Thu, 02 Dec 2010 00:10:11 GMT

    [ https://issues.apache.org/jira/browse/UIMA-1875?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12965933#action_12965933
] 

Jörn Kottmann commented on UIMA-1875:
-------------------------------------

Handled the overlap problem a little different now, the tag is placed depending on the location
of the annotation.
It tries to always center the tag below the annotation, but that is not possible if the annotation
is in the begin
or end of line and an overhang occurs.
The overhang is now allowed to be two chars, if the tag is longer its trimmed down.

We have to see how this works in practice maybe the overhang should be reduced to be only
one char. It also works much nicer if the text itself uses a bigger font size than the default
12. 
Hopefully the default font size will be increased by default to 14 in the release after 2.3.1.

Its still possible to have overlapping tags, I hope that can be optimized a little more in
the future.

Please report your experience back, so we can improve this feature.

> ability to visualize and quickly update/add values to primitive features
> ------------------------------------------------------------------------
>
>                 Key: UIMA-1875
>                 URL: https://issues.apache.org/jira/browse/UIMA-1875
>             Project: UIMA
>          Issue Type: New Feature
>          Components: CasEditor
>            Reporter: Philip Ogren
>            Assignee: Jörn Kottmann
>         Attachments: CasEditor-TagDrawingStrategy.tiff, CasEditor-TagDrawingStrategyOverlap.tiff
>
>
> I spent a bit of time evaluating the CAS Editor recently and have the following suggestion.
 It is common to have annotation tasks in which adding a primitive value to a annotation feature
happens frequently.  Here's one common annotation task - part-of-speech tagging.  Usually,
the way this task is performed is a part-of-speech tagger is run on some data and a part-of-speech
tag is added as a string value to a feature of a token type.  The annotator's task is then
to look at the part-of-speech tags and make sure they look right and fix the ones that aren't.
 However, the only way to see the part-of-speech tag is by clicking on the token annotation
in the text and view the value of the feature in the editor view.  This makes the tool really
unusable for this annotation task.  What would be really nice is to be able to display the
part-of-speech tags above or below the tokens so that the linguist can scan the sentence with
its tags and quickly find the errors.  
> There are a number of other annotation tasks that have similar requirements.  For example,
named entities usually have category labels which would be nice to display.  Word sense disambiguation
data is also similar.  

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


Mime
View raw message