File system based data portability
Ideas, OPSN, Social Network May 21st, 2008
In my recent Chinese blog I wrote a post about some ideas on the data portability.
I gave an example on the traditional PC, the users data can be accessed from the different applications, and they can also be accessed from the file system. Image what will happen if you can not have the file system level access?! It will be terrible. We backup, copy, share most of our those document simply based on file, NOT the application which create and view them!
Look at today’s web world, it’s a different story:
User can only access their data which store on the service provider’s servers from the specific interface, they can’t access from file level! That’s why the data portability problem happens!
Unix system have a very good design in file system, and it intend to map any devices into a node in the file system, so access a printer or scanner is just like open a file for read/write. I think a network file system alike data portability design is a one-fit-all solution, which can be very helpful for data portability.
If I use this idea into my hobby project OPSN (Open Portable Social Network), things become much simpler: what if the import/export of contacts are just as simple as copy some vCard files? We can implement a simple filter to scan/parse microformats and simulate them as files.
And what if some third party build a universal file system adaptor service? It build a virtual file system, and mapping flickr, facebook, gmail, … all as paths of the virtual FS? It looks like a awesome idea! Maybe that’s a direction to the true next generation Internet OS! (Flickfs now can mount flickr account to the Linux’s file system, we need more!)
–
I read some blogs, they have similar ideas:
http://avc.blogs.com/a_vc/2006/11/data_portabilit.html
He thought of the data could be stored in 3rd party storage system, e.g. S3, or simply a personal NAS. e.g. buffalo link station.
http://webworkerdaily.com/2008/04/25/data-portability/
Data portability & file system. Exactly same idea as mine.
http://deanlandolt.com/archives/205
He have similar idea and he is starting a projected call "socialFS"!
About
Very interesting — yeah, we’re definitely barking up similar trees. I particularly like the visuals…
As for how to get something like this off the ground, I can’t help but think the actual file system aspect of it is secondary — at least from the standpoint of portability. Coming up with a model and a means to reappropriate a lot of your data is job one — exposing it through various APIs (an important one of which being FUSE or something like it) can only come after.
I am thinking of WebDAV, how do you think?
I am not familiar with WebDav, juast learning and thinking if it’s a good idea.
We have worked in the similar direction - http://code.google.com/p/semap/. It’s an RDF based ‘filesystem’ developed in RESTful style. Unfortunately, we stopped the development and many our ideas exists in Russian only
Dulanov, your project sounds very interesting I will learn from it in the next few weeks. Fortunately, I have native russian speaker in the team, I will ask them for help to understand your idea.
Why you stoopped the development?
Honi soit look closer buy cytotec then announced estivities.