[okfn-discuss] Data that service users consider private, but want to publish from the service

Fri Jul 18 12:44:27 UTC 2008

On 17/07/08 10:43, Francis Irving wrote:
> How broad is as "personal in nature" for the purposes of part 1. of
> the Open Software Services Definition? It seems to only include secret
> data - for example personal emails, or my private calendar.
> 
> There seems to me a category of data which is personal/private, but
> where the data is not secret. It is just the user only wants it
> using in ways they control. For example:
> 
> 1) My own blogs posts are data that I'm happy to be publically viewable,
> but could be sufficiently personal I don't want to give them away as
> open data (e.g. I might want to tell a personal anecdote in context on
> my website, but not want it to appear as lines in a broadway musical)

Sure in that case you wouldn't be using an 'open' license (or even a 
'non-commercial' type license for that matter).

> 2) I might want to share photos of myself on a photo sharing site for
> my friends and anybody in the world to look at there, but for privacy
> reasons not want them to be taken and used as the next face of
> l'Oreal.

Sure and again you'd then want to steer clear of an open license.

> Which leads to the more general question - if the users of a service 
> want to manage their own data with the service, but they want that
> data to be neither open nor secret (but somewhere inbetween), can the
> service still be OSSD?

Hmmm. My first instinct is to say yes. However, i'm concerned this might 
allow end-runs around the intent of the definition. For example, what 
about YouTube. If I remember the ToS there basically give YT the right 
to redistribute but no-one else.

Nevertheless, this an example where the material is user's material and 
so I think we should stick to the line that has been drawn. It is only 
if it data used as part of the service and *not* supplied by users that 
one has an obligation to make it openly available (remember that user 
data must be made available for users to extract in an automatable and 
complete manner though).

> So for example I could fork gitorious (an open source code hosting
> service http://gitorious.org/projects/gitorious), and allow secret
> repositories, with a simple interface for getting your data out of
> your private repository (git pull :). Call this fictional site
> private-gitorious. I think it would comply with the OSSD.

Yes I think it would.

> Then suppose I fork gitorious, but let people host in public non-open
> source software on it, but for which the source code is available
> (e.g. Microsoft shared source style). Call this shared-gitorious. I
> would still make the source code of gitorious open (I'd have to, by
> the Affero GPL), and obviously people would be able to get their own
> data. Indeed, anybody would be able to get anybody's data. Just
> *other* people wouldn't be able to reuse it.

Sure but I think we need to be clear here what is the service and what 
is the content. The service here is the gitorious. This service exists 
to *host* user material. That material is not actually part of the 
service but what the service supports (in this regard Wikipedia is 
different -- there the content/data is an essential part of the 'service')

> In many ways private-gitorious is more "closed" than shared-gitorious,
> yet it is, I think, more clearly OSSD compliant.

I think both are compliant.

> Apologies if you've covered this already.

No, these are excellent questions. I'd be interested to hear others' 
take on these matters.

~rufus