Thomas Meehan gave a presentation to the Cambridge Cataloguing Group today and explained the Library of Congress Bibframe initiative which builds on Tim Berners-Lee's concept of a semantic web where the information in the World Wide Web is arranged in orderly machine readable fields. To be exact what he actually said was:
Libraries currently create bibliographical records for books based on the RDA standard and using MARC21 code. These two standards do not sit easily together and when RDA was created it was envisaged that one day MARC21 would be supplanted by another standard more suited to the semantic web concept, in short Bibframe. The Library of Congress put a lot of effort into spreading the RDA standard across world libraries and now they are feeling brave enough to start experimenting with the next step, Bibframe.
Cambridge University Library owns over three million books. Each title published has, in theory, has its own Bibliographical record and on each of these three million or so records there is an author and a subject listed. In a good library management system you can click on the name of the author on one document and display a list of all the titles created by that author. However, the argument goes, this is a bit of a waste of data because if you had a separate record for that unique author then you would not need to re-write that name twenty three times, once in each of the records for the twenty three titles that that author wrote. In a relational database each piece of data is stored as an entity with a relationship linking it to another piece of data. The entities in Bibframe are called URIs and are a bit like a URL except they aren't necessarily live.
A librarian asked Thomas how a cataloguer's work would be affected by all this, and the answer was surprising because he showed an example of an online data input entry form and it looked very similar to any other online data entry form that one might care to name at the moment in any of the more intuitive cataloguing clients for systems such as Unicorn or Heritage.
The bad news that there are lots of rivals inventing different versions of Bibframe and each creating their own URIs which are therefore not unique at all, which defeats the purpose of harmonisation and unification. In theory you would only need one authority record for an author and all the other records in the world could link to that definitive record. So it is really important to agree on a standard. Moreover the surprising thing is that many of the proposed Bibframe models do not particularly reflect the FRBR models on which RDA was created.
I raised the issue that there was already an International Standards Organisation and Book Industry standards that ensure that data is standardised to be managed through EAN's and UPC's so I thought it was important that we work together with these agencies to ensure that our library practice fitted in with international data standards from other organisations who are all facing the same sort of issues. Across the internet all those who sell things will be wanting to ensure that their product data is as machine readable as possible, which means even more rival versions, so unification and consultation are important issues here or we will simply be adding to the tower of Babel effect so insidious in the software industry.
Often we librarians get tied up in our own little library world and don't realise that we are reinventing the wheel: for instance library management systems are just a specialised form of the databases that all companies have to manage their product data, and our query process is very similar to the query process that all companies face and usually manage through customer relationship management systems (CRM) such as Zoho.
"I have a dream for the Web [in which computers] become capable of analyzing all the data on the Web – the content, links, and transactions between people and computers. A "Semantic Web", which makes this possible, has yet to emerge, but when it does, the day-to-day mechanisms of trade, bureaucracy and our daily lives will be handled by machines talking to machines. The "intelligent agents" people have touted for ages will finally materialize".[8]
 |
| Image by Zepheira, under contract from the U.S. Library of Congress [Public domain], via Wikimedia Commons |
Cambridge University Library owns over three million books. Each title published has, in theory, has its own Bibliographical record and on each of these three million or so records there is an author and a subject listed. In a good library management system you can click on the name of the author on one document and display a list of all the titles created by that author. However, the argument goes, this is a bit of a waste of data because if you had a separate record for that unique author then you would not need to re-write that name twenty three times, once in each of the records for the twenty three titles that that author wrote. In a relational database each piece of data is stored as an entity with a relationship linking it to another piece of data. The entities in Bibframe are called URIs and are a bit like a URL except they aren't necessarily live.
A librarian asked Thomas how a cataloguer's work would be affected by all this, and the answer was surprising because he showed an example of an online data input entry form and it looked very similar to any other online data entry form that one might care to name at the moment in any of the more intuitive cataloguing clients for systems such as Unicorn or Heritage.
The bad news that there are lots of rivals inventing different versions of Bibframe and each creating their own URIs which are therefore not unique at all, which defeats the purpose of harmonisation and unification. In theory you would only need one authority record for an author and all the other records in the world could link to that definitive record. So it is really important to agree on a standard. Moreover the surprising thing is that many of the proposed Bibframe models do not particularly reflect the FRBR models on which RDA was created.
I raised the issue that there was already an International Standards Organisation and Book Industry standards that ensure that data is standardised to be managed through EAN's and UPC's so I thought it was important that we work together with these agencies to ensure that our library practice fitted in with international data standards from other organisations who are all facing the same sort of issues. Across the internet all those who sell things will be wanting to ensure that their product data is as machine readable as possible, which means even more rival versions, so unification and consultation are important issues here or we will simply be adding to the tower of Babel effect so insidious in the software industry.
Often we librarians get tied up in our own little library world and don't realise that we are reinventing the wheel: for instance library management systems are just a specialised form of the databases that all companies have to manage their product data, and our query process is very similar to the query process that all companies face and usually manage through customer relationship management systems (CRM) such as Zoho.
Comments