[odc-discuss] odc-discuss "Making Your Data Open: A Guide"

Wed May 20 09:56:03 UTC 2009

2009/5/19 Andrea Wiggins <akwiggins at gmail.com>:
> This looks like a nice start - I think the guide would benefit from
> more specific how-to steps/examples under "So How Can I Make My Data
> Open?" The text that is currently there provides arguments about why
> to do it rather than examples of how to go about doing it, which would
> probably be more helpful.

I think you're absolutely right. We should probably drop the whole of
the "Why Does Openness and Licensing Matter?" section.

> E.g., for #1 - What mechanisms can people use to make data publicly
> available? This would be a good place to link to repositories that
> host open data contributions, and information about setting up an API
> or web service, for example. It might also be a good idea to mention
> that people should evaluate the appropriateness of making a given data
> set open (e.g. human subjects data, some types qualitative data) and
> provide links to other venues for sharing restricted-access data. This
> provides a useful support to anyone who wants to share their data but
> for ethical or other reasons cannot make it an open data set.

Good idea. Do you have any good suggestions here. We might want to
make this a link out to e.g. a wiki page as I can imagine it will grow
and change over time.

> Also, I think it's important to say that good metadata and data
> documentation are crucial to making open data reusable, since it
> shouldn't be assumed that readers will be familiar with data
> preservation best practices. I would go so far as to say that if there
> are no usable metadata or documentation, then the data are hardly open
> at all; from my experience, a data set without metadata is practically
> useless for reuse. Pointers to models or resources for how to document
> a data set would be very helpful.

Another good point. My one thought though here is whether we want to
keep this document more focused on the "legal" howto side of things.
After all ODC are focused on the licensing/dedication side of things.
One solution would be to have a short section with a pointer to
better/bigger resources elsewhere.

> For #2 - It might be handy to add info about providing citation
> information along with the license for those licenses which require
> attribution. If you want attribution, then making attribution
> information easily available is important.

Absolutely. I think we should definitely add to the PDDL example a
mention that they can also apply a set of community norms (such as the
one we have already drafted) -- as that's where the attribution info
would go in that case.

Rufus