Few words of wisdom on playing around with ODC.
1. While setting up the batches in all the tools(ODC, Import, Recognition, Commit) for a particular use case, always remember to keep the Batch Prefix same. Else you will keep wondering I’ve indexed the batch , but why it is not getting committed. I’ve successfully imported a batch, but why its not showing up in the Commit Server. So, make sure you double check the prefixes that you are setting up for a particular use case.
2. While setting up the commit driver for a commit profile, you need to provide the complete link to your content server, like http://myserver.com/cs/idcplg. Remember to add /idcplg at the end, else you will not be able to retrieve the elements in next screen completely, like the security groups , or accounts information.
3. Don’t delete the folders manually that are created while processing the scanning and indexing batches. The associated batches still show up while selecting them for indexing , and I was unable to remove them.
4. Use zonal OCR with discretion, it’s not 100% perfect, it’s a suggestive technology, as I would like to call it, which needs verification for every attribute recognized via OCR. Even Oracle documentation says, the values “might” appear in the metadata fields, and not, values “will” appear in the metadata fields. You get the difference?
5. If you are setting up bar-code detection in your scan profile , you have to keep the “File Type” as “Tiff” and “Color” as “Black and White”, easier for the system to read the barcode with just black and white , rather than a full scale RGB values may be.
6. While setting up the Capture db and Indexing db, keep a watch on the path that you are setting up. Either default , in the poc environment, or a shared one in an Enterprise setup. Have a knowledge of these database tables to understand how ODC keeps this info and for debugging as well.
7. While setting up the Import server, if you want the output format as Tiff/Jpeg, you need a printer app, so that the import server starts printing the docs/xls/ppts as an image, and if there is any issue with the printer app, you need to do the process again. So, if you are not trying to perform any image specific operations, keep the output format same as what you are inputting one.
8. Plan for your Enterprise wide ODC workstations setup, as every component has a defined work, and need access to the batch folders created at some or other point in the ODC process.
And last, but the most important one.
Do your homework. Put the process flow on paper before you actually start configuring the ODC and other components.
UCM Chronicles 