Posting to NGC4LIB


I think we are beginning to see eye-to-eye again. I agree that library methods are mainly 19th-century, our processes are 1970s, and the *public face* of what we are building is divorced from the needs of our users. I don’t think FRBR or RDA solves or even seriously discusses any of this, and why I think it’s all going down the wrong road.

So, the power that is represented by our authority files should be unleashed into… what? Topic maps? OK, I’m all for it! How about putting it all into a wiki or lots of wikis and let the world play? For example, the Mark Twain record is really useful and could be made much better, but not in its present format, and it’s locked away in a closed database anyway. Let these things out for general experimentation, see what works and what doesn’t, and go from there. This is impossible to do while the data remains walled off however. If it were up to me, all of it would be available in a couple of hours, but…

I doubt many of you have deep knowledge of what is possible these days through various forms of AI, clustered facets, semantic models and various degrees of formal logic over reasonably fuzzy structures of knowledge management. Of course it’s easy to counter this by saying “well, show me, I haven’t seen it.” Well of course you haven’t ; no librarian specialist has joined up with a systems expert to create it yet (or, at least, I haven’t seen it :).

I can only hope something like this happens soon. Still, while I am sure I don’t know everything about AI, I don’t think that matters to the point I am trying to make, which is: currently, it can’t do the job that we, as librarians, feel needs to be done. I may be wrong, but I haven’t seen it yet, and I haven’t seen anything even close. If you want us to believe you, it must be proven to us, even in a prototype. If we bring up the example of how we have controlled Mark Twain’s name and ask if that control (not the method but the control) can be replicated in some way with AI, don’t try to convince us that retaining that control is not important. Show us what the machines can do, and it is then our job to decide if it’s good enough or not. But I am not just going to believe it.

And yes, it is precisely librarians and I think catalogers who should decide because they know what to look for, and not leave it to an inexperienced and naïve public. This is like some of the spiritualists who would claim they could talk to the dead and make tables walk around the room. They could fool lots of people, including professors with vast experience, but they couldn’t fool Harry Houdini. It’s the specialists in describing and organizing information (catalogers) who know the pitfalls.

Still, the *only* reason we do have the control we currently have over our materials is because of our adherence to the standards. If we throw off the standards, we throw off the control. Therefore, it must be clearly thought out.

> Welcome to my world! This is what bibliographic control is all about and it is not simple.

Well, is that because of the tools and methods chosen, or because it’s an actual difficult area? The reason I’m asking here is for example the dreaded way of having birth and death years (sometimes, or too often optional, or wrong) as part of the name as an identifier for entities which you have to match and merge in order to make sense out of (and most of the time get wrong). In the past these identifiers were the best one could come up with because there were no computers, no decentralized means of resolving them, no authority control across different bodies. However, now we do. See where this is going?

I agree with a lot of this. For example, I think the Wikipedia disambiguation page is far superior to our methods, but still, the actual task itself of “disambiguation” is difficult, although I agree it can and must be made much simpler. We must always keep the standards in focus though; if we were not involved in a cooperative task, we would not have to follow standards shared with others, and we would be free to do anything we want. But shared standards involve other responsibilities.

Should the standards change? Absolutely, and RDA purports to do that, but in my opinion, it’s just more of the some using other vocabulary, and why I initiated the Cooperative Cataloging Rules.