Manual review of data

From reBiND Documentation
Revision as of 19:05, 4 November 2014 by LornaMorris (talk | contribs) (Manual Review and Corrections)
Jump to: navigation, search

Manual Review and Corrections

The results from the automated correction can be manually reviewed and edited via the web interface. It is especially importanat to review the errors and warnings. Errors and warnings could be caused by technical issues or with the content of the document itself. More minor changes are flagged as 'info' messages. If there are any errors or warnings it is advised to have both, the contributing scientist and the content administrator go through the review together. Errors in the document could be fixed by modifying the correction configuration and re-running the automated correction with a new configuration file.

There is an online XML editor included in the eXist release, called eXide, an online demo is available at the eXist homepage. The eXide editor is a modification of the online source code editor ACE - Could9 Editor. With eXide it is possible to directly edit documents stored in the database, including features like syntax highlighting or code folding.

The eXide editor that comes packaged with eXist has been modified to create the reBiND editor.

Reviewing corrections in the reBiND Editor

A screenshot showing the results of the correction on the reBiND_Puffinus.xml data file is shown below:

When the reviewer clicks on any of the items in the list, the editor will directly jump to the element that was changed within the document. A modified GUI could also allow to only view changes of a certain type (class) or done by a certain module. It could also allow to flag changes as reviewed and hide reviewed changes from the list.

Though this might be a good way for the Content Administrator and the Technical Administrator to review the changes, there is still the problem that the Contributing Scientist is confronted with the XML document and required to work with it. So at some point in the future a better interface for the Contributing Scientist to work with the data might be desirable (e.g. the use of automatically generated web forms to edit the data), but for the general infrastructure described in this text, the online XML editor is sufficient.

It is not necessary to review all the changes at once. The document can be stored at any time and the review process can be resumed at a later point. So it is possible that after the correction one of the administrators reviews the changes which are caused by technical issues or the XML format used, leaving only changes which are related to the data. Then the Contributing Scientist can review the remaining changes.