I've worked to deprecate marc-hash (what tends to be referred to as "Bill
Dueber's JSON format") in favor of Ross's marc-in-json. To the best of my
knowledge, there is marc-in-json support for ruby (current ruby-marc), PHP
(current File_MARC), marc4j (currently in trunk, soon to be released, I
think), and perl (MARC::Record in the next release). I think that covers
all the major players except the IndexData yaz- stuff.
[Galen, any word on that next release of the perl module?]
I, at least, already use marc-in-json in production (It's a great way to
store MARC in solr). It would be great if folks would have the confidence
to use it, at least as a single-record format. I think for wider adoption
we'll need to all have either (a) json pull-parsers to read in a file that
contains an array of marc-in-json objects, or (b) a decision to use
newline-delimited-json (or some other record-delimiter), so folks can put
more than one of these in a file and be able to get them out without
running out of memory.
-Bill-
On Thu, Dec 1, 2011 at 9:11 AM, Ross Singer <[log in to unmask]> wrote:
> Ed, I think this would be great. Obviously, there's zero
> standardization around MARC/JSON (Andrew Houghton has come the
> "closest" by writing up the most RFC-y proposal:
> http://www.oclc.org/developer/content/marc-json-draft-2010-03-11).
>
> I generally fall more in the camp of "working code wins", though,
> which, solely on the basis of MARC parser support, would put my
> proposal in front. In the end, I don't think it matters which style
> is adopted; it's an interchange format, any one of them works (and
> they all, including Bill Dueber's) has their pluses and minuses. The
> more important thing is that we pick -one- and go with it so we can
> use it with some confidence.
>
> While we're on the subject, if there are any other serializations of
> MARC that people are legitimately interested in (TurboMARC, for
> example:
> https://www.indexdata.com/blog/2010/05/turbomarc-faster-xml-marc-records)
> and wish ruby-marc supported, let me know.
>
> Thanks,
> -Ross.
>
> On Thu, Dec 1, 2011 at 5:57 AM, Ed Summers <[log in to unmask]> wrote:
> > Martin Czygan recently added JSON support to pymarc [1]. Before this
> > gets rolled into a release I was wondering if it might make sense to
> > bring the implementation in line with Ross Singer's proposed JSON
> > serialization for MARC [2]. After quickly looking around it seems to
> > be what got implemented in ruby-marc [3] and PHP's File_MARC [4]. It
> > also looked like there was a MARC::Record branch [5] for doing
> > something similar, but I'm not sure if that has been released yet.
> >
> > It seems like a no-brainer to bring it in line, but I thought I'd ask
> > since I haven't been following the conversation closely.
> >
> > //Ed
> >
> > [1]
> https://github.com/edsu/pymarc/commit/245ea6d7bceaec7215abe788d61a0b34a6cd849e
> > [2]
> http://dilettantes.code4lib.org/blog/2010/09/a-proposal-to-serialize-marc-in-json/
> > [3]
> https://github.com/ruby-marc/ruby-marc/blob/master/lib/marc/record.rb#L227
> > [4]
> http://pear.php.net/package/File_MARC/docs/latest/File_MARC/File_MARC_Record.html#methodtoJSON
> > [5]
> http://marcpm.git.sourceforge.net/git/gitweb.cgi?p=marcpm/marcpm;a=shortlog;h=refs/heads/marc-json
>
--
Bill Dueber
Library Systems Programmer
University of Michigan Library
|
