October 28, 2005
It's tonight. And at the last minute, even though I have something else this weekend, Anita, danah and Kaliya convinced me to go. So.. I'm going. Can't wait.
Been working on a tagging solution today involving Identity, Creative Common's licensing, and Tags for two different types of tagging in personal publishing (blogging). Drummond Reed and I have been working on this periodically for about 6 months, and finally figured out the security and privacy aspects at our meetings this past week, though we'd figured out the syntax and usefulness of it last April and June. Kaliya Hamlin has also been helping a bit. Looking forward to talking with folks about it to see if it makes sense. It's called iTags because it includes identity in the tags.
The idea is that a user could tag an object (photo, video, sound file, text or an entire blog post), where the tag, and the object, would then go out through the RSS feed or be spidered, with some additional information that doesn't now exist in tags. That tag and object would include the user's identity, the licensing for that object (presumably people would use this more for rich media objects than for just a blog post, as most blogs already have licensing generally for text on the blog) if needed, and the tag. It would remove the requirement for a tag to be coupled with the originating URL (blog post URL) because identity would be inside it. It would allow individual CC licensing which rich media producers want to do sometimes, in order to differentiate the rich media object from the rest of their blog, which may have different licensing.
This solution also addresses these user desires:
1. create tags without links that are still visible in the blog post
2. create tags that are not visible but still trusted (like categories which are not visible but still included in systems like Technorati's tag system, where of the 100% of tags, 95% are actually category entries and only 5% are actual tags).
It still allows users to create tags that do have explicit links if the user wants them, as is already being done. It's backwards compatible with Technorati, but allows additional functionality to meet these other goals.
Wiki for this stuff is here. Please feel free to add to the wiki or comment here about this. Would love to know what you think.
Posted by Mary Hodder at October 28, 2005 04:21 PM
My preference is to have localized identity system that the owner of the rich media can implement on their individual server. Rather than centralized dictionary authorities. In a distributed system it seems the best implementation is as atomically individual distributed as possible.
I wrote my impressions about iTags in a post on Clipperz.net. The title is: "iTags, well done Mary Hodder!"
In short I think iTags is a very sound approach to a more useful and powerful tagging system.
Hi Enric, Thanks for your comments. We will definitely talk about what you suggest, to see how it fits with our user goals.
Marco, Thanks very much. I did see your post, and appreciate your support.
If either of you have ideas about how to make this better, I encourage you to go to the wiki and post them. Marco, I'd like to get your RSS suggestions posted as they are very helpful.
Moving beyond our other discussion to the meat of the proposal, I was curious as to how this differs from http://www.Microformats.org/ and why you did not consider working with them on this (or id there a shortcoming I dont know about). It would seem that with the mutliple pieces of data they handle with their approach, it might be better to work XRI into there to handle this problem - particularly if you mean it to be produced by a tool/utility as you have stated elsewhere rather than human hand as I suggest with TagSpaces.
The suggested implementation seems like this addresses keywords originating from a creator rather than tagging, which in contrast, is done personally as well as collectively. But perhaps there is a simple semantic misunderstanding between how I am understanding your problem and how I view tags. My view on tags is largely driven by the human folksonomy approach rather than a technical approach.
I see what you are describing as a Media Subscription Descriptor MicroFormat that contains information about the creation of the object, rather than its identification by the creator AND others via tagging. As I have said previously, I see the tag as one mutually agreeable string of text that connects the searcher with the identifier - the enabler of an "identify, find and refind" value exchange. In the sense you have described it, it seems more like an XRI spec than its own thing, though I do see the brilliance of leveraging the rel tag to make it backwards compatible in order to prevent technologies that rely on them from becoming obsolete.
I am not as technically astute with regards to envelopes and encapsulation in regards to RSS of podcasts, but assume you are - what iare the current shortcomings you are addressing with this? - is it just the lack of CC license, identity and tags? What of this is handled in the atom spec? If not, why is it not also being proposed there at the source, or included in RSS 2.1?
My understanding is that the rel="tag" is a microformat already, so since iTags use that format, it is now compatible with that microformat.
Also, I understand that microformats are set up to be spidered, where as we want this to work everywhere, including RSS feeds which have far more reach than microformats. Right now microformats only work for companies that have a spider looking at blogs.
But that said, of course we'll put the format out and if they are interested in it, great. If not, the version of iTags with links should work right now, according to their tag directions. It's the versions of iTags without links that will not be backwards compatible with Technorati/microformats and that will need to be addressed with Technorati later. We do intend to work with everyone and try to be compatible with everyone, including Technorati.
I recently listened to Robert Cringely Nerd TV interview with Robert Kahn, http://www.pbs.org/cringely/nerdtv/player/?show=012&ext=mp3 . At the end of the show around 55 minutes in, Robert Kahn describes his newest work started around 1995, "Digital Object Architecture". You probably know Robert Kahn is one of the inventors of the internet defining TCP/IP. "Digital Object Architecture" ( http://www.cnri.reston.va.us/doa.html , http://tinyurl.com/rfsvb ) focuses on the development of an infrastructure of services that provide access to distributed and secure digital objects. Kahn states that the technology is already in place and being used by digital journals. It appears to provide the structure for digital identity and seems worth looking into.