On Wed, Jan 7, 2009 at 1:15 PM, Thommy M. <thommy.m.malmstrom@gmail.com>wrote:
Joe Atzberger wrote:
This error isn't related to ASCII vs. Unicode. You have a mandatory field in your framework (942$c) that isn't populated by your import. Populate it manually to continue.
I'm very well aware of that. The word "anyway" was to indicate that I had two (2) problems/questions. I both have a problem with unicode _and_ an error message. I can also figure out that the mandatory field is not filled in, however there is no way to fill it in.
So question 1 is; How to set this field or how to make it not mandatory?
This is controlled by your MARC frameworks.
The problems that Koha does have with Unicode are in large part problems that any MARC21. Our French colleagues (using UNIMARC) certainly would not be using Koha if it could only handle ASCII.
I very well know that Koha can manage unicode, I only want to be provocative to get some attention to the fact that there is a big lack in the documentation on how to set it up correctly.
Asking for counterexamples as though you doubted such installations exist is not the way to get documentation. It is the way to get counterexamples.
Question 2 is; How can I make Koha to use unicode?
What configuration parameters should I set? Can I send some list of current parameters to describe my settings so that someone can tell me what I've set wrong? Is there any current documentation on how to get Koha unicode aware?
Check that you have correctly defined the "encoding" field for your z39.50 targets. Most MARC21 targets will be MARC8 encoding, so don't be lured by the otherwise attractive UTF8 option for those. To make sure the z target is actually representing and transmitting the data correctly, try connecting on the command-line with yaz-client or a different stand-alone z39 client. Otherwise, the INSTALL docs and especially the encoding_scratchpad wiki have the most complete documentation for Unicode happiness. --Joe