WHY
I do. Any post I make I have squared away. This allows me to make my site the definitive collection of my data. It also allows google to index it. I have control over my own content and if for some reason a third party site wants to exert control I still have my stuff. Unlike slashdot who didn't allow any tools to save posts. Hence I lost from about 1996 to 2002's worth of comments. [0]
HOW
There are 2 ways of looking at this. You can either generate you "stuff" from a single point [0] at your machine, save it and pump it out OR use the web apps as clients and suck the data back via RSS, Atom, JSON etc.
I've been doing a bit of both. Pushing stuff from my blog engine after I've cached it and now I'm beginning to suck up the various websites I frequent. So for pushing out I have:
- OUT: flickr (blog, tags, images) , twitter (snippets), hackerid (hackernews data), links (various links I save including links to hackernews & export to delicious)
- IN: hackernews (all posts every 15m, friends),
- IN TODO: wordy (words I use), spock (tags), colourlovers (colours), librarything (my library), amazon (new books), lastfm (what I'm currently listening to), delicious (new links I find), twitter (friends), flickr (friends, processed images, text, tags)
Now as you can see that's a lot of data. Some of the things I'm finding:
- It's easier to push than pull if you want an accurate copy because you save before you export
- pulling data means you don't have to write firstly the interface to capture the data and simply call RSS
- not every site has an API or good RSS feed
- Linking data together is not easy except by time though you could try to match by friend (ie: friend is on twitter, flickr, hackernews)
- displaying the data effectively is difficult simply because of the volume and complexity of it. A good example of how to do this is http://friendfeed.com Clear, simple and pretty much allows for good reading.
I'm now at a stumbling block with a templating engine I'm using and so I'm pretty keen to just extract the data as Atom, RSS and JSON as individual feeds or a mashed feed by date and write a Javascript based website to avoid having to deal with heavy weight blog engines. Let the data go free and see how people use it.
The key thing to realise is when you are compiling your data timestamp it at DB level (if you are using one) ISO1606 format, maybe add a tag layer over the top so you can get the benefit of tagging across data layers.
It's turning out to be an interesting project.
[0] http://goonmail.customer.netspace.net.au/2005DEC131709.html
Standard time in the Netherlands was exactly 19
minutes and 32.13 seconds ahead of UTC by law
from 1909-05-01 through 1937-06-30.
The bit I use is a bastardisation of "1985-04-12T23:20:50.52Z". A shorter version as a string, say "20070202T1422" stripping out the hyphens, colons and Zulu.
time-secfrac = "." 1*DIGIT
time-numoffset = ("+" / "-") time-hour ":" time-minute
time-offset = "Z" / time-numoffset
partial-time = time-hour ":" time-minute ":" time-second
[time-secfrac]
full-date = date-fullyear "-" date-month "-" date-mday
full-time = partial-time time-offset
date-time = full-date "T" full-time
20070202T1422 - but close enough for me in 1 timezone and limited space for display. I have a double use as both a (8601 hack) and as a human readable title. Adding extra "-" and "Z" makes it harder to read and the accuracy for seconds is simply not required. Trade-offs I'm willing to make.
Our goal is to provide a place where people can store and access all their online files. We want to be the Online File System. We're heading to the "cloud" already and it's becoming increasingly difficult for us to manage our own files. My photos are locked in Flickr, my docs in Zoho, etc.
We want to break down the "data silos" and provide a single interface to all your "cloud storage" as well as provide a single API for developers to code against to access user's data in meaningful and interesting ways.
We're taking a multi-pronged approach to this. One prong (currently launched) is to allow you to upload your files through our API or our web interface and we'll store it (using Amazon's S3 for reliability and scalability). You can access the files and their metadata through the UI or through the API. You can tag them, share them, make them public, etc.
Our next step is to provide many backing-stores. That is, import your files from Flickr/Zoho, etc. We'll keep things sync'd up, so that Openomy becomes your Online File System; storing the metadata of everything you do online. Then, the UI becomes the equivalent of Windows Explorer or Finder, and the API becomes the equivalent of the single file system API provided to you as a desktop developer.
We think it's going to be, at the very least, extremely fun to build. :)
(bias alert: I'm deeply involved in the project I'm about to mention)
A few years ago I realized that I'd either lost data due to hard-drive crashes or mistakes and wanted some sort of safe place to put my data. We then started a project (http://allmydata.org) which provides for a decentralized, secure storage fabric. You can create your own grid if you'd like, create one with friends, or use our managed grid (for a service fee).
I think that Amazon S3, Nirvanix, and others are starting to provide nice basic services in this area as well, but I do think there is a lot of room for retaining and saving information that is gently locked into web services (Flickr, blogs, LinkedIn, etc.).
This allows me to put my data "out there" among all the computers/friends that I share my space with, without publishing the information for the whole world to see. So I have ease-of-access, backups, online/offline access, etc without having to go public. Plus it tightly integrates with my operating system and office software, as opposed to my having to write a bunch of hand-coded stuff.
It certainly says something when you can get better information from Wikipedia than from Microsoft about one of their own products.