Recent experiences in using shark as a Workflow

1. There is a problem with hibernate in shark 1.0.
When I run it using hibernate, errors are always returned:
Net. SF. hibernate. jdbcexception: cocould not initialize collection: [org. enhydra. Shark. usergroup. hibernategroup. Users # sale manager]

Caused by: Java. SQL. sqlexception: [Microsoft] [sqlserver 2000 driver for JDBC] [sqlserver] An error occurred while converting the Data Type nvarchar to numeric.

Then I looked at its hibernategroup. HBM. XML, and I couldn't do it.

Either version 1.0 or CVS is incorrect.

2. Shark modules are very independent. This time, I need to replace my own user table and role table. The result shows that if I change the usergroup module, it will be OK, and other modules will not be moved.

3. Get historical information without the operator. I want to get the historical operation records of a process. I can get historical activities using the shark API, but I cannot get operators of historical activities. I have to get them directly from the background database.

4. It is best to directly call the liberary API without using its CORBA interface. I feel that it is faster, mainly because my application is B/S.