Wednesday, February 16, 2011

RE: rdacontent terms - dataset

Posting to RDA-L

Jonathan Rochkind wrote:
<snip>
On 2/15/2011 10:34 AM, Weinheimer Jim wrote:
>
> I am being real. The plain text format of MarcEdit *absolutely cannot* do the same as MARCXML. I'll prove this right now. Browsers are built to work with XML, so right now, this second, any webmaster can work on the fly with XML using nothing more than a browser. They need no other tools.
That's not really true. Most browsers don't give you any tools for 'working with' XML, although some (but not all) of them will display it with nice syntax-aware coloring.
</snip>
Sorry to contradict you, but I have done this myself multiple times. Here is a discussion of it: http://www.herongyang.com/XML/Browser-XML-File-Browser.html. Anybody can work with XML and XSLTs with a browser, and in fact I have had to do it because I did not have access to the expensive XMLSpy, which verifies your XML.


But I think I finally understand where the disagreement lies. For example, you mentioned:
<snip>
Nonetheless, I see no reason to think getting either "MarcEdit text format" OR MarcXML to be processed natively by our ILSs would be an improvement for us at all. If that's what's being suggested? I'm not sure what IS being suggested.
</snip>
No, I am not suggesting ILSs. If everything were based on everyone searching separate library ILSs, everything would be fine but that is no longer the case. The internet is growing through means of mashups and apis. This is an absolute fact. This is a wonderful development and extremely powerful. But what does this mean exactly? I think the best explanation is this short video from ZDNet that I suggest everyone watch. http://www.youtube.com/watch?v=ZRcP2CZ8DS8.

In my opinion (and not only mine), this is the world we must enter, whether we want to or not. How do you enter this world? By creating Web Services. In order just to start to do this, you must use XML, since this is the language. It is not ISO2709. What are some examples?

Let us imagine that a scholarly group wants to build a site about baroque architecture (this is true since I know some of them). One thing they should be able to do is to make automatic queries behind the scenes (i.e. web services) to bring together all kinds of information to create some new tool of use to their community. They can--right now, today--use Google Maps, Google Books, Amazon, Yahoo, dbpedia, the Internet Archive ... Here are just a few of them available now: http://www.programmableweb.com/apis/directory. In here, we can see that there is a system called BookBump http://www.bookbump.com/ that uses a number of apis, including the LC SRU API http://www.loc.gov/standards/sru/, which is based on providing records in.... XML.

Unfortunately, there still doesn't seem to be a Google Scholar API, which could be one of the most important apis for our community.

If we do not enter this world based on APIs and web services, I fear that we will be left behind completely. The general public will never even know about our records. We must let our data enter and interoperate with other apis in ways we cannot foresee right now. We also cannot expect that everyone will consciously click on the links to our catalogs and search them, because they just won't. Besides, people want to use our information in genuinely unique ways they never could have before.

This is why I feel so strongly that sticking with ISO2709 for transferring records hurts us terribly. The longer we remain in that ISO2709 straight jacket, the less we can enter the world where everything is happening. There are other reasons, too, but in the world of mashups, we cannot assume that people will come to our ILSs, especially when they will be able to use the Google Books api or the LibraryThing api or the Internet Archive api. There must be some kind(s) of Library api.

No comments:

Post a Comment