Quoting Tim Spalding <[email protected]>: > But could you really walk back from OL to MARC? It seems to me > OL-to-MARC conversion is a lossy, one-way conversion. You could get > some sort of placeholder record, like libraries sometimes have for a > pre-publication book, but not much more. > > Shall we go through examples, or shall we just agree on that?
I actually did some work in that direction, so it's not just theory. No, of course not, from OL you do not get a "complete" MARC record. However... 1) where there was a MARC record as input, it is possible to get back to it. Right now that's only available via the UI, (go to history and it's usually the bottom item) but with some code it could be available through the API. (I'd be interested to know if anyone can import into an ILS from the display of the MARC record which is there. ) There are disadvantages, in that you wouldn't get any additions/corrections in the OL record. Doing that latter is feasible, just more code. Here's a display: LEADER: 00716cam 22001931 4500 001 48000245 003 DLC 005 20041206175439.0 008 731023s1948 maub 000 1 eng 010 $a 48000245 040 $aDLC$cDLC$dDLC 050 00 $aPZ3.L81146$bRai$aPS3523.O246 100 1 $aLockridge, Ross,$d1914-1948. 245 10 $aRaintree County$b... which had no boundaries in time and space, where lurked musical and strange names and mythical and lost peoples, and which was itself only a name musical and strange. 260 $aBoston,$bHoughton Mifflin Co.,$c1948. 300 $axiv, 1066 p.$bmaps.$c22 cm. 651 0 $aIndiana$xFiction. 655 7 $aHistorical fiction.$2gsafd 655 7 $aBildungsromane.$2gsafd Here's a raw marc record: 00716cam 22001931 4500001001300000003000400013005001700017008004100034010001700075040001800092050003300110100003300143245019200176260004200368300003200410651002200442655003100464655002700495 48000245 DLC20041206175439.0731023s1948 maub 000 1 eng a 48000245 aDLCcDLCdDLC00aPZ3.L81146bRaiaPS3523.O2461 aLockridge, Ross,d1914-1948.10aRaintree Countyb... which had no boundaries in time and space, where lurked musical and strange names and mythical and lost peoples, and which was itself only a name musical and strange. aBoston,bHoughton Mifflin Co.,c1948. axiv, 1066 p.bmaps.c22 cm. 0aIndianaxFiction. 7aHistorical fiction.2gsafd 7aBildungsromane.2gsafd These are linked from this page: http://openlibrary.org/books/OL6026352M/Raintree_County?m=history 2) I believe that some of the records came in as MARCXML, some as "binary MARC". So possibly there would need to be cleanup-recreation so that everything would be in the same format. Not sure, though. 3) the input files, including the LC Books files (still missing one year, 2008 I believe, but up to date with the weekly files since then), are stored at the Internet Archive. If someone wants to make a database of MARC records out of those, it should be possible. Some of the non-LC library sources turned out to produce some pretty error-ful records (I do not want to name names in public). I would advise against using those. (Contact me offline) Here's where to find them: http://www.archive.org/details/ol_data 4) OL does not store data in "library cataloging" format. In particular, it was decided to reverse the names from last, first to first last. This isn't reversible, at least not accurately. With code, one could restore those X00 names, but there would need to be a strong motivation. I would like to see the LC name authority identifier added to OL records as just another identifier for names, but I don't think that's high up on anyone's list. But in general, no, OL is not producing data that fits into the MARC/AACR bag. kc > > Tim > _______________________________________________ > Ol-tech mailing list > [email protected] > http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech > To unsubscribe from this mailing list, send email to > [email protected] > -- Karen Coyle [email protected] http://kcoyle.net ph: 1-510-540-7596 m: 1-510-435-8234 skype: kcoylenet _______________________________________________ Ol-tech mailing list [email protected] http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech To unsubscribe from this mailing list, send email to [email protected]
