print() method of the JTable accepts a PrintRequestAttributeSet which is actually a part of Java PrintService API. It allows the application developer to add any attributes that are given under 'javax.print.attribute.standard'. The issue here is how many of them are really supported by the underlying platform / print service. Currently a print service on an average supports 10-12 attributes and rest of them are ignored. It varies from one print service to the other and varies from platform to platform.

On Solaris and Linux, IPP is supported in tiger via CUPS and on CUPS, one can not expect a print service to support the same set of attributes what it supported on Non-cups. So the best way is to check what is supported and what is not before passing the attributes to the JTable print method. So the JTable.print() method must clearly specify that it is the responsibility of the application developer to check using the print service methods whether an attribute is supported by the print service or not before letting Jtable use the attribute.

Currently the API doc for print() just provides a link to PrintRequestAttributeSet which is not at all sufficient. It is very likely that the user would use some of the attributes that are not supported by the underlying print service and it is hard to expect every user of the APIs to know the intricacies of Java Print Service APIs.

Moreover, as I mentioned above, when he runs an application on a different platform or with a different print service or on a CUPS vs Non-CUPS system, it is very likely that he would see lot of differences in the attribute support which he might assume as a failure. Such complications could be avoided if the attribute support is verified upfront in the application. So it is better to document these points in the JTable.print() method just to make the users aware of print service/platform dependencies for attribute support.