different types of (remote) storage

[Melvin Carvalho]

Some of you may be interested in this vocabulary that Tim Berners-Lee is working on.  It provides a way to do discovery from a users via storage to what he calls "workspaces".  Workspaces are where you can save data remotely for use by client side unhosted apps.

It documents the idea of different times of (remote) storage that a user can have, and how to add "rel" links for discovery:

:Storage  a s:Class; s:label "storage";
    A storage is a space of URIs in which you have access to data.

:PublicStorage  a s:Class; s:subClassOf :Storage; s:label "public storage";
    A public storage is a space of URIs in which you have access to data,
    and all data is accessible to anyone without control.

:PersonalStorage  a s:Class; s:subClassOf :Storage; s:label "personal storage";
    A personal storage is a space of URIs in which you and only you have access to data,
    you cannot give access to anyone else.

:ControlledStorage  a s:Class; s:subClassOf :Storage; s:label "access controlled storage";
    A  storage is a space of URIs in which you can individually control for each resource
    who has access to it.

http://www.w3.org/ns/pim/space

If unhosted libraries such as RS could one day parse this kind of rel link for discovery I think that would be great as it would be double (or more) the amount of users, developers, testers and apps

In any case I hope it may be some interesting "food for thought"

[Nick Jennings]

Thanks for sharing this Melvin, I'd like to make use of existing work in this area wherever possible, though in this case I think taking on an entirely different markup language in Turtle, is taking on quite a bit of new complexity for our limited resources.

I'd like to see the ability to embed storage location information inside a webpage using RDFa, though. This could be something added to remoteStorage.js without necessarily needing to be written into the spec, since it's just about finding the remote storage location, and not about interacting with it. You would either enter either a user-address (in which case WebFinger is used) or a URL (in which case we attempt to look for a known set of RDFa properties). I think this would be a great way to open up some additional flexibility.

At the moment, I think we simply don't have the manpower to really consider taking on the feature, but it's something I often think about. I'm all for expanding the ways in which we can interoperate. It would be great to have a number of standardized ways for defining your personal storage endpoints supported, so you were not forced to setup a webfinger endpoint if you just wanted to embed that information in your website.

It would also be nice to see a JSON-LD version of the above Turtle definition (for my own better understanding). I've been looking around on the web for this lately (also for WebAccessControl defined in JSON-LD) but haven't seen much. From what I've read, I guess hard-core Linked Data folks find XML and JSON to be too limiting to define some of the structures possible in Turtle. However, I must admit, it's difficult to justify making the effort to take on another full-fledged data formatting language just for these use-cases. I'm very happy right now in JSON world, and like JSON-LD quite a bit...  (and RDFa is only a small addition to my existing comfort zone of HTML)   :)

Cheers
Nick

--
 
---
You received this message because you are subscribed to the Google Groups "Unhosted Web Apps" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unhosted+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.