http://bugs.koha.org/cgi-bin/bugzilla/show_bug.cgi?id=997 Summary: No MARC subfield sequence specification or subfield repeatablity Product: Koha Version: unspecified Platform: All OS/Version: All Status: NEW Severity: critical Priority: P2 Component: MARC AssignedTo: paul.poulain@free.fr ReportedBy: koha@alinto.com QAContact: koha-bugs@lists.sourceforge.net GENERAL PROBLEM The Koha MARC data model has a design mistake that is more than a mere implementation bug and results in data loss for affected records. The cataloguing forms have no means for subfield sequence to be specified correctly by the cataloguer when the sequence requried by the item to be catalogued differs from the Koha default subfield order. Even when the correct subfield order is imported from an external record, the Koha MARC framework has no means to preserve the order of subfields in data storage when the order differs from the Koha default order. Furthermore, while the cataloguing forms have a means to enter repeated fields, they have no means to enter repeated subfields within the same field when required by the item to be catalogued. This problem would never be seen in very many common and simple cases but it this problem exists in Koha for many fields when the item requires a different data model than Koha uses. This is a terrible problem for intent of MARC records to be shared between systems, where the already corrupt data pool would become corrupted by these records. The problem may be most acute for subject fields where the complete form of the subject field with subdivisions should match the form found on other remote systems to properly locate other items with the same complete subject. Even where externally created records are imported into Koha, such as the former records of an institution prior to adopting Koha, the data will be corrupted and information will be lost. The information conveyed by the subfield sequence is lost for the problematic records. No automated process could recover the lost sequence information for most cases without referencing the original uncorrupted records; or some nonexistent intelligent system actually understanding the meaning of the subfield content and reordering the subfields according to real world knowledge of how they should properly be ordered. PROBLEMATIC EXAMPLES Some examples of the most problematic effects this problem has on the creation and interpretation of records follow. Many other examples may be taken from the same and other fields but these seem especially problematic. 260 ##$aParis :$bGauthier-Villars ;$aChicago :$bUniversity of Chicago Press,$c1955. Multiple publishers cannot be entered using the cataloguing forms. An externally catalogued record imported into Koha would be transformed as 260 ##$aChicago :$aParis :$bGauthier-Villars ;$bUniversity of Chicago Press,$c1955. Note that the place of publication and publisher are not only not adjacent but the correct place of publication would no longer match the correct publisher if reordered from this point. 490 1#$aDepartment of State publication ;$v10846.$aDepartment and Foreign Service series ;$v12128 . Multiple series titles cannot be entered using the cataloging forms. An externally catalogued record imported into Koha would be transformed as 490 1#$aDepartment and Foreign Service series ;$aDepartment of State publication ;$v10846.$v12128 . Note that the series title and volume number are not only not adjacent but the correct series title would no longer match the correct volume if reordered from this point. 505 00$tQuark models /$rJ. Rosner --$tIntroduction to gauge theories of the strong, weak, and electromagnetic interactions /$rC. Quigg --$tDeep inelastic leptognnucleon scattering /$rD.H. Perkins --$tJet phenomena /$rM. Jacob --$tAn accelerator design study /$rR.R. Wilson --$tLectures in accelerator theory /$rM. Month. Formatted contents notes cannot be entered using the cataloguing forms. .An externally catalogued record imported into Koha would be transformed as 505 00$rC. Quigg --$rD.H. Perkins --$rJ. Rosner --$rM. Jacob --$rM. Month.$rR.R. Wilson --$tAn accelerator design study /$tDeep inelastic leptognnucleon scattering /$tIntroduction to gauge theories of the strong, weak, and electromagnetic interactions /$tJet phenomena /$tLectures in accelerator theory /$tQuark models / . Note that the title and statement of responsibility are not only not adjacent but the correct title would no longer match the correct statement of responsibility if reordered from this point. 650 #0$aArchitecture$zIllinois$zChicago$xHistory$vPictorial works. Repeated subject subdivisions cannot be entered using the cataloguing forms. An externally catalogued record imported into Koha would be transformed as 650 #0$aArchitecture$vPictorial works.$xHistory$zChicago$zIllinois . Note that if reordered from this point, an automated system would need to know the subsidiary relationship of Chicago to Illinois for correctly reordering the subfields. SOLUTION APPROACHES It should be obvious why many many cataloguing systems have one free entry text field for all subfields. I the guided data entry approach that Koha uses for cataloguing preferable for its self validating effects. A means to include repeated subfields should be incorporated into the cataloguing forms similar to the means for repeated fields. The subfield to be entered at a particular point in the subfield list should be alterable from a drop down selection list for those fields that may have a variant subfield sequence and those records that may not have a varient subfield. The actual coded values for the subfield should have the position of that subfield in the sequence of subfields for that selection list for maintaining the subfield sequence in the Koha database. No implementation should prevent the use of a recommended or required value list or thesaurus for the subfield contents. If the page needs to be rewritten when the cataloguer alters the default subfield se sequence of subfields to accomidate a value list or thesaurus in the cataloguing form corresponding to the altered subfield sequence, then the page needs to be rewritten. The default cataloguing subfield sequence should provide for the most usually correct subfield order as a default that corresponds to the different usage of different fields. A parallel data structure to the existing data structure can be introduced to preserve the subfield sequence if there is no easy means to augment the existing data structure with subfield sequence information. At least newly catalogued and newly imported or reimported records from an uncorrupted source should have their subfield sequence preserved in the data structure after a suitable implementation is developed to correct this problem. ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact.