Hello, Am 24.07.2007 um 18:01 schrieb Graham Triggs:
> On Tue, 2007-07-24 at 11:12 -0400, Robert Tansley wrote: >> ... However, IDs may depend both on the object type and >> related objects -- e.g. bitstream IDs may include the item ID as a >> path component, or the version number etc -- and the ID scheme >> itself. > > IDs will (? should!) always be free of the context of the process of > assignment. Am 24.07.2007 um 17:32 schrieb Ekaterina Pechekhonova: > Do you think it's appropriate for external IDs to support "internal" > system hierarchy and version numbers ? (I am not arguing just > wondering) I am arguing and I agree with Grahams view that IDs should not code any metadata of the object identified. I hope my story is not to lenghty because it is only loosely related to the subject, but for me it fits in here. The former programmer of our project has changed the handle manager to support child naming authorities. During my ongoing effort to upgrade to 1.4x, I had to read and understand the code and was thinking a bit how Handles are created and stored in the database. We had to do this change to convince contributing parties to put their data into our repository. They were not clear about the way they could get back their metadata when our project is loosing funding (which is the case right now BTW). The export alone was not the whole answer as the persistent identifier was still dependent on our system remaining online. Our changes are not the kind of quality that we can send them in as a patch right now and following this discussion I am not sure that it will get that status before getting obsolete. You will find some problem while visiting our site right away, the most obvious of them being communities constituing child naming authorities not being entities of their own child naming authorities but of the parent autho- rity, thus remaining the only dependent item when moved to another repository. Aside from that, why do I strongly believe that metadata and identifier should not be mixed? First, this could lead people to "forge" URL because they believe they know who the scheme is. This is happening right now with some crawlers on our site and I am not yet sure that we dont provide them ourselves with the wrong link, but I cant see how we do this right now (my logs are not complete and I will follow this issue). Second, more important, the current model, where IDs are a sequence is not the best solution as well. This sequence gives metainformation and might lead to contradictory con- clusions. The lower IDs tells you that this item has been created before another item. But, you can change the date through the submission interface. This might be relevant in some cases where a submitter has to prove credibly that his work has been published before or past a certain date. As admin of the repo you will be in charge to make the origin of the difference between date order and ID sequence of the two tiems be understood. I would object to help the lawyers with that. I am interested in the integrity of my repo but not in others to use this fact at their will. I think, the current model should be changed to some fixed length hash value of sufficient length, but I dont know how to do that with the means of the database. It should be done by the db because the ID is a primary key of the table where it is stored. Also, such a change need to support preexisting handles as well, probably requiring an additional column in the table or other schema mods. Bye, Christian ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ DSpace-tech mailing list DSpace-tech@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dspace-tech