Tuesday, January 19, 2016

The biggest IT changes in the last 5 years : The fragmentation of notifications

I remember the days of "push". I remember Microsoft CDF (Channel Definition format). I remember, RSS in its original Netscape form and all the forms that followed it. I remember ATOM...I remember the feed readers. I remember thinking "This is great. I subscribe to stuff that is all over the place on the web and it comes to me when there are changes. I don't go to it."

But as social media mega-hubs started to emerge (Facebook, Twitter, LinkedIn etc.), this concept of a site/hub-independent notification infrastructure started to fragment.

Today, I find myself drawn into Facebook for Facebook updates, Twitter for Twitter updates, LinkedIn for LinkedIn updates, Youtube for Youtube updates....I am back going to stuff rather than have that stuff come to me.

At some stage here I am going to have to invest the time to find a mega-aggregator that can go aggregate the already partially aggregated stuff from all the hubs I have a presense on.

Rather than look for updates from the old-school concept of "web-site", our modern day aggregators need to be able to pull updates from hubs like Facebook, Twitter etc. which are themselves, obviously, providing aggregation.

The image this conjures into my mind is of a hierarchical "roll up" where each level of the hierarchy aggregates the content from its children which may themselves be aggregates.

The hierarchical/recursive picture has a lot of power obviously but I do wonder if it has the unfortunate side-effect of facilitating the emergence of web-gateway models for hubs. I.e. models in which the resources behind the gateway are not themselves, referenceable via URLs. We end up with no option but to "walk" the main nodes to do aggregation.

I remember a quote from Tim Berners Lee where he said something along the lines of "the great thing about hypertext is that it subverts hierarchy."

Perhaps, the mega-hubs model of the modern web subverts hypertext?

Monday, January 18, 2016

The biggest IT changes in the last 5 years : domain names ain't what they used to be

The scramble for good domain names appears to be a thing of the past. A couple of factors at work I think.

Firstly, there is obviously a limited supply of {foo}.com and {bar}.com. Out of necessity, "subdivided domain names" seems to be getting more and more popular. e.g. {foo}.{hub}.com and {bar}.{hub}.com are on the same domain name.  So too are {hub}.com/{foo} and {hub}.com/{bar}.

This has worked out well for those who provide hub-like web presence e.g. facebook, github.com, bandcamp.com etc.

Secondly, browser address windows have morphed into query expressions, often powered by Google under the hood. Even if I know I am looking for an entity {foo} that I know owns {foo}.com, I will often just type into the search bar and let the search engine do the rest.

Extending DNS with new domains like .club etc. only pushes the problem down the road a bit.

I am reminded very much of addresses of locations in the real world. Number 12 Example Avenue may start out as one address but it may decide to sub-divide and rent/sell apartments at that address. Now you have suite 123,Number 12, Example Avenue....

Nothing new in the world. DNS is like Manhattan. All the horizontal real estate of DNS is taken. The only want to get a piece of it now is to grab a piece of an existing address. DNS has entered its "high rise" era.