Tuesday, January 10, 2006

2006 - The Year of the Unique ID?

While people like me are publishing automatic produced predictions for 2006 others like David Weinberger are taking a more serious look at what might be happening this year.

Journal Of Hyperlinked organization - December 5, 2005: "Last year, it was Web 2.0 and tagging. This year, it's going to be unique IDs (UIDs), and for the same reason that Web 2.0 and tagging matter: The Web is going miscellaneous." <SNIP> "Which is why I think UIDs are more likely than global UIDs. Competing groups will come up with their own schemes, perhaps labeling incommensurate objects, and then as the need becomes pressing, we'll map the systems together, however awkwardly. That may require a lot of footnotes, but experience has shown (anyone remember SGML?) that we're better off having relatively local groups succeed at ID'ing objects and then knitting them together than waiting for the World Council on Numbering Things to come up with a global standard. That hasn't happened since Adam and Eve, and even they made an arbitrary decision to names classes of things ('Let's call them 'dogs'') and not particulars ('Let's name that thing 'Rover' and that thing 'Fido'')."

The full article is quite an interesting read, and I think he is right that Unique ID's will be come increasingly important in the next few years of internet history. But this will only happen as this ID's become more (a) useful and (b) visible.
I will talk about (b) first:
(b) Currently there are a whole lot of ID's that we do not see, traditionally the model has been to "hide the ugly id's" so that we don't scare away our users. The same has been true of the likes of RSS and other "computer read" information. Traditionally we have tried to hide all the techie stuff but the rss movement has shown that by exposing some of the raw data, people are suddenly able to do some very cool stuff.
Examples of openly available UID's are already occurring today, want to link to a flickr photo?
It has a unique id : http://www.flickr.com/photos/jambe/59292611/
a url on del.icio.us?

These ID's are specific to the individual sites involved and don't help you find photo's that are essentially the same, or find to url's that point to the same website. But they are the first step in what David is talking about. They are also both URL's which you have to say are the most common unique ID's in use today. A lot of UID systems build upon URL's for this reason. This practice doesn't fit very well with the completely random style UID's David W seems to be advocating.

(a) Usefulness - Unique ID's have always been useful, they are used in almost every database, printed on cars, used on the internet, and in networking. However so far there has been little emphasis on exposing these numbers to users. I think that visibility and usefulness will go hand in hand, as UID's become exposed, they will suddenly become more useful, and the more useful they are the more that are likely to be made available.

I also think it is worth noting that small groups creating Unique ID's creates a naturally hierarchical structure (which seems to go against Dr. Weinberger's Internet philosophy - he is a philosopher BTW)

Example hierarchy (This hierarchy doesn't really exist, all the names are all made up):
root-> Microsoft's UID's -> BLOG_ID
                                 -> PERSON_ID
         -> Google's UID's -> UNINVERSAL_INFORMATION_ID
                                     -> ADSENSE_PERSON_ID

Now I guess If everyone started using Microsofts PERSON_ID system (may repeat these are all fictional names) the hierarchical structure would be largely meaningless, but it does show that hierarchies do exist, and they do have uses. If you haven't read about the tagging/tree debate, before you definitely should... I'm still trying to form an opinion.

No comments: