[openbiblio-dev] bibjson -request for feedback

Jim Pitman pitman at stat.Berkeley.EDU
Wed Feb 22 00:11:49 UTC 2012

Mark MacGillivray <mark.macgillivray at okfn.org> wrote:

> The datasets I have copied, with confirmation from you, are at:
> http://my.cottagelabs.com/~ichi/bibkn_datasets/

OK, thanks. I'll take a look. I would like to see these beneath bibkn.org however. Please 
let me know how I can help arrange what is put there.  Mainly I want to repair broken
links as best we can.

> > I would like to see stable urls beneath
> > bibjson.org for the old spec from the bkn grant,  and for newer versions since then, and a
> > default spec address which always goes to the most current spec.
> > I think it is important to have some discipline and stability about versions of the spec.
> > If there is some way to say there is no spec, by adopting other standards for our purposes, then
> > great. But even then, we will need some documentation of how we are adopting those
> > standards for our purposes. I presume that documentation will become operationally the
> > latest BibJSON spec, right?

> The page at bibjson.org is the latest convention on the use of
> BibJSON. The current proposal is to update that page removing the
> "namespace" usage, and with links to JSON-LD spec for detail of LD usage.

OK. This still seems too casual to me. I would like to see something more rigorous,
e.g. numbered versions that can be referred to. It should be possible to see the evolution
of the spec, not just the most current version. This seems very basic to me. 
Can you point to other successful projects that are so casual about their data model?
I would like to hear others views on this.  

> > And even then, the earlier specs are artifacts of some historical importance, which are
> > already referred to on the web. It looks bad for our projects to have links e.g. from
> > our STM paper to sites under our control yielding 404's or blank pages. Lets repair this
> > situation quickly as best we can before we put out the blog posts about transfer of the bkn name and assets to okf control, since those posts should refer to a presentable
> > bibkn site with an adequate set of files to represent outcomes of the former bkn project.
> Within scope of discussion re. bibjson convention, bibkn.org could be
> pointed at bibjson.org to bring people following old links to the
> latest bibjson information.

OK, maybe as an immediate fix, redirect bibkn.org to bibjson.org until we figure out
what to do with bibkn.org. But I would not like to see this for long. And it is not compatible
with our blog posts on bkn embedding in OBWG which are now ready to go except for lack of
a bibkn site.

> > I have similar concerns about the bibserver code.  Shouldn't this be versioned and
> > archived in some way as we proceed?
> Yes, it is in a git repository and available at http://github.com/okfn/bibserver

Is there a notion of numbered releases?
What is the plan there?

many thanks


More information about the openbiblio-dev mailing list