Thanks Ian, I read your powerpoint before. I am trying to do similar thing, I have tried to crawl the xrecord and i am confused with the code generated to represent Multi byte character. (most of our record use Chinese character). III XML generate special encoding, something like {21357} and i am really having problem to convert those character back to Unicode. And after those exportation, are there any way to automate the process? For example, catalog librarian created a new record, what is the proper way to auto-update XML record without using Create List function? Wayne 2010/10/19 Kyle Banerjee <[log in to unmask]> > On Mon, Oct 18, 2010 at 7:28 AM, Wayne Lam <[log in to unmask]> wrote: > > > Any people have experience in exporting bib data in Innovative Interface > > to sort of XML db? How is it going to be done actually? > > and What could be the best way of doing this? > > > > What's the use case you have in mind? Also, when say bib data, does that > mean item data is not needed? > > IMO, the easiest way to get bib data out of III is just doing a MARC dump > of > the bib files. The process is clean, you can convert the data to whatever > format you like, and you're golden > > If you are only interested in certain fields, it might be easier to use > delimited output from Create Lists -- particularly if how the fields are > indexed in III is important to you. > > The other major alternatives include making sure xrecord is enabled and > then > just picking all the records up in XML. Unless the III specific fields > (e.g. > bcodes, suppression codes, etc) or what internal III tags are associated > with each MARC field are important for your application, I recommend > against > this approach. III generates the second suckiest XML I've seen in my life > so > it's a PITA to work with (the State of Oregon CMS generates XML with a > level > of suckiness that will never be topped). > > You can also crawl the public OPAC and parsing the textual representation > of > the MARC bib record. If you don't have admin rights on the machine, this is > not a horrible way to go since you don't need to worry about waiting for > people to help you, dealing with Create List files that are only a tiny > fraction the size of the entire DB, etc. III bib records are numbered > sequentially, so it's easy enough to guess an entire database and pull up > every record in the webopac by bib control number. > > kyle >