Hi Kevin, I'll chime in here as well, just to give you one more e-mail to read. I think there are perhaps three questions here:
(1) What is your motivation for doing this? You mentioned Endnote as the driving factor for your project. If I may, I would actually suggest that that should probably *not* be your focus. As Ed mentioned previously, not many researchers use Endnote's own search facility to download citations. More often when searching databases these days, they will use the database's export facility to download the records directly from the web into Endnote. In fact, Thompson, the company that makes Endnote, recommends this other approach over their own search facility, which is clumsy and requires users to update connection files periodically as databases change. If you would like to allow Popline users to download citations, you can, as Ed mentioned, simply allow them to select and export the records via the web. Most databases that support this feature use the RIS format, which most bibliographic citation software in turn support. In that way, you can offer this service to people using any citation software, not just Endnote. http://www.refman.com/support/risformat_intro.asp I would suggest that a better motivation for making Popline searchable via Z39.50 or HTTP/XML is to accommodate the increasing number of libraries who are using metasearch systems. Over the next three to five years you'll see more and more researchers using metasearch systems to access bibliographic content, and for Popline to remain relevant and useful, you'll definitely want to allow metasearch systems to search your site. All the major systems can support Z39.50 or HTTP/XML. (2) What protocol do you want to use? The choice here is really either Z39.50 or HTTP. Note that MARC and Dublin Core are metadata schema, and are independent of the choice of protocol. You can send XML over Z39.50 or MARC over HTTP. To me, Z39.50 is not an attractive option -- it's an aging, arcane protocol, and would require a good deal of work on your end to set-up. If you already had a robust Z39.50 infrastructure, I'd say keep it. But if you're looking to put something in place now, HTTP offers a much simpler and easier means to make your records available. You're already using a web server for the Popline web interface, so you're half-way there. If you use HTTP, then XML becomes the best markup language for sending back the records. Again, this is independent of a metadata schema. There is an XML version of MARC, called MARC-XML, and XML versions of Dublin Core, and pretty much everything else. (3) What metadata schema do you want to use? Regardless of the protocol, you'll want to choose a metadata schema to export the records in. Here the question turns back to the first one: What is your motivation? If you are looking to position Popline for the increasing number of libraries using metasearch, then I think you need to choose a format that is (1) widely used, (2) robust enough to support the creation of OpenURLs, which many metasearch systems will create to allow users to locate a local copy or submit an interlibrary loan request for an item. The last point is key. Popline is a great resource and we recommend it to our users here. But without any OpenURL capabilities, it is currently isolated from our other system. Our users want to know how they can actually get their hands on the books and journal articles they find in Popline, and simply won't do a manual search these days given the ease-of-use of SFX and other link resolvers. For the two reasons above, I would highly recommend that you use MARC-XML, and that you implement it in an SRU set-up, as Ed mentioned. All of the major article databases, from ERIC to PsycInfo, provide records in MARC format, and that is the orientation of most metasearch software. Doing that in SRU with MARC-XML makes a lot of sense. http://www.loc.gov/z3950/agency/zing/srw/sru.html I think all of the metasearch systems on the market today can handle SRU/SRW, and sending the records back in MARC-XML makes it much easier for libraries to get Popline up and running quickly. If you choose Dublin Core or a proprietary schema that's more work on our end. OpenSearch does not support a robust enough set of metadata to create an OpenURL, and I think that makes it a very poor choice. MARC (in an XML representation or otherwise) can easily handle journal articles. It wasn't really intended for that, but neither was Dublin Core, or really any other metadata schema. So there you go. :-) Just my $0.2. I think it's great you are looking into this. It will make Popline much more accessible and useful to a lot of people! --Dave -----Original Message----- From: KEVIN ZEMBOWER [mailto:[EMAIL PROTECTED] Sent: Thursday, September 01, 2005 8:47 AM To: perl4lib@perl.org Subject: General question: Should MARC be used for periodicals and journals? I have a general, perhaps philosophical, question. I'm trying to make my collection of reference material on reproductive health called POPLINE (check it out at www.popline.org) available through what I thought were Z39.50 servers. I was basing this assumption on what I thought was a Z39.50 client in a program called Endnote that the researchers in my organization use to search other medical databases, such as PubMed (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi). I always just assumed that this was a Z39.50 client and responding with MARC records, until a librarian I'm working with to understand MARC told me that MARC's main purpose is to catalog monographs, and that it's not particularly well suited to represent articles in journals and periodicals. Further research on the PubMed site now leads me to believe that Endnote must have written a client that allows it to search PubMed on port 80 and get answers in XML format. I'm going to ask this question of the PubMed support list, but can anyone on this list tell me off the top of their head if this is correct or not? If this is correct, I'm faced with a problem. About a quarter of the entries in POPLINE are monographs and the rest are articles appearing in periodicals. Should I try to force this into a MARC format and use a Z39.50 server to allow access? Or, should I write an HTTP API and allow searches on port 80 and return information in XML? I'd like to be as similar to any standard or de facto standard, such as PubMed, as possible, to aid in adoption and ease of use. Is there a well-accepted standard for articles in journals similar to MARC if MARC shouldn't be used? Should I, or even can I, write a Z39.50 server that returns a result in MARC if it was a monograph, or in XML if it was an article? Finally, I know this list is specifically about the use of perl in libraries, and I'm taking quite a liberty asking these questions here. Is there another, more broadly focused list that I should be asking questions like this on? Thank you for all your advice and suggestions. -Kevin Zembower ----- E. Kevin Zembower Internet Systems Group manager Johns Hopkins University Bloomberg School of Public Health Center for Communications Programs 111 Market Place, Suite 310 Baltimore, MD 21202 410-659-6139
