I use Twitter more than anything else nowadays. More than any other social media – facebook, linkedin, Google Reader, Wikis, Social bookmarks or even podcasts. I use all these, but the links actually seem to generate from the tweets I read.
Twitter is my quick environmental scanner. Twitter stream brings me nuggets of news, opinions, humor and recommendations. Why it is even my medium for mini-conversations and a source of discovering some cool people and products.
This wish list is really me thinking aloud. Some of them may be Twitter enhancements. Others may come from Twitter clients.
1. I get follow notifications by email. So I would love to see an email plug-in (Gmail preferably) that provides a tag cloud of the recent tweets by the person who follows me. And an interface to follow right there and add them to one of my lists would be a bonus.
2. I would love to consume tweets by email (that way I have one universal client) but many people may not really like this. If it comes to email, I can do some clever filing with rules. I am sure one exists. I may just need to find one that works.
3. I would love to have a twitter client or intermediary with a rule engine support. A little more powerful than the rules you find in Gmail today. There are lots of interesting possibilities here.
4. I would love to have curated.by as an email client too. In fact, Google should borrow Curated by chromium extension’s tagging functionality for Gmail. I know I am digressing a bit here.
5. I would love to get tweets in a different priority order, sorted by a retweet count (from my friends). This should be alternative to a chronological order. Just another view.
6. I would love to see a visual social graph on twitter – a static connection graph and a dynamic one based on active conversations.
7. I definitely would like a better search engine. Since I am dreaming, I may as well ask for one with better contextual and semantic search. This keyword search does not do it for me.
I have a few more (a template tweet for conferences for example), but will keep it as material for another post.
What are your wish list items?
Nice article. Long since I thought twitter is just a internet SMS but today it become newspaper for me! anything everything today become twitter. Facebook, Orkut etc.. are secondary.
If you are windows user you can try Seesmic Desktop 2. i am using Ubuntu so its not supported in linux :(. but i like Tweetdeck!
Nice article. Like to see more from you 🙂
Thanks. I use Tweetdeck on Windows, Twidroyd on my Android, Twitter email client and some times Amplify and Curated.by. But I would love to do it from my email client and do all these other things as specific functions.
You will hear from me. I am one of those noisy channels 🙂
I think there is one more think that application like twitter needs and that is some sort of automatic categorization, classifying people into groups based upon the kind of content they generate. Some people talk about their life chores others talk about books and reading material and many talks about technology.
This would help users in finding/following users generating information of their interest and also to people like you to clean up your list of followed users who are generating content loosing your interest.
I know this can be done be only through machine learning technique. And this makes me ask you a question: Do you know any application that does something of this kind categorization based on content?
Thanks for taking the time to add your input. I think categorization is a complex problem. It needs include several factors – the context (was it a retweet, answer to another tweet, part of a larger discussion indicated by hashtag, event tweet). The context is not isolated to a single tweet (even though it may be, in some cases) but to the tweeter, the time/date, currency of information etc.
It should be possible to do some simple analysis. Keyword filtering is next to useless (the one employed by many search tools). Here are a few factors I would consider in doing a twitter rank.
1. Implicit/explicitly stated bias of the Tweeter with respect to the subject
4. Some kind of tweet rank based on a mix of quality and quantity of followers
5. context (as I mentioned earlier) both local and a broader one
There may be others. If you procure sufficient information, some machine learning techniques may work. That is yet to be proved. Most of the current attempts at gauging the sentiment or mining information are not there yet ( I don’t profess to know them all).
I think curation is a great first step. When I curate a topic, I am careful to choose relevant topics. However, since I am limited in the number of tweets I can read, my coverage of the subject area may be questionable.
Just rating tweeters is not good enough. You need to rate the tweets. There are some interesting possibilities here.