Twitter is responding to critics with an admirable degree of humility.
I don’t think it’s a case of armchair quarterbacking, more a case of Twitter being in the “show me” penalty box that comes with having immense promise and goodwill, and simply not delivering on a key expectation. Twitter has, as a company, had one real responsibility put on them by their community, make the backend stable. There has been no loud outcry about client side improvements or richer functionality, none of the stuff that other companies have to deal with, just make it work reliably.
Simply put, Twitter fails too frequently and only recently has the company begun seriously addressing these issues in a public fashion. I still love the service but I’m more interested in seeing improvements, less interested in long winded explanations about why their reliability sucks. It they want to post these details, I’ll listen but when it comes to actual improvement in service, show me.
Part of the impetus for this public discussion extends from the sense that Twitter isn’t addressing our architectural flaws. When users see downtime, slowness, and instability of the sort that we’ve exhibited this week, they assume that our engineering progress must be stagnant. With the Twitter team working on these issues on and off for over a year, surely downtime should be a thing of the past by now, right? Shouldn’t we be able to just “throw more machines at it”?
[From Twitter Technology Blog: Twittering About Architecture]
Wow, it’s been a busy week here at NewsGator central, today we announced a pretty cool feature in our client products for content recommendations called Collaborative Filtering.
We are pretty hot on this idea of filtering content and providing recommendations based on the attention you give to your subscribed content. This includes not just clicking on a post in a feed to read it, but also clipping, tagging, and sharing content, all of which come together in our Activity Scoring.
What we are doing with Sense Array is actually pretty blind to the actual content, it uses attention streams as simulated ratings on top of explicit thumbs up/down ratings the user initiates on suggested content. We have been working with mSpoke on a rich categorization capability that will provide Wikipedia-style indexing of content, and you will being seeing that in the products in the near term.
So where is all this going? Well first things first, we need to finish integrating this with our desktop client apps so that you have the same recommendations in FeedDemon, NetNewsWire, and your Outlook inbox, and it’s all powered by the online engine. This also suggests that we can make this available to third party app developers through our API.
Secondly, we are taking advantage of our widget services for content syndication as a logical delivery point for relevance features that provide suggested content on top of what our media partners are delivering. Ironically, this is a core part of what we offered with our branded reader product that led to the development of the widget business, so in a sense this is coming full circle to providing publishers with rich services on top of RSS, just as we do for consumers.
Lastly, I think it’s fair to say that the emergence of exciting new RSS consumption applications and the embedding of RSS services into applications means that there is a movement to provide a new user experience for content consumption that does not rely on a subscribe and consume chronologically model. This next generation paradigm is what captures my interest and I think we can deliver it.