Search to Activity

This is the idea that instead of URLS one could just type something like Mark, and a bunch of potions would pop up. These options would include things like, contact methods (email, im, phone), sites, feeds, projects etc.. I think this is most interesting because right now there are a lot of good solutions that seem to stay independent. For instance, if I type something into Google I will find a lot of things. If I type something into Address Book I may find some things, come things I want but perhaps some of the things I want will not yet be back in Address book. Similarly, if I wanted to email someone rather fast I would essentially first have to create a new email. If I wanted to instead of emailing them look at their LinkedIn page I would have to either know their LinkedIn URL or first search or go to my contacts list in LinkedIn. 

My argument is that all this data, like the URLs of a bunch of my friends LinkedIn pages, and their email addresses and their websites and their blogs are all stored in places that are accessible already (many are stored simply in the browsers URL auto complete. So, there should be a thing that checks who a page or account is about and then associates them with it. So if i start typing there name in or a service in I get either a semantically useful list of their sites or a list of people I know using that service respectively. Just to give a completely obvious example, when I type Will (someone I studied with) in the right place I should see an option to instantly email him, I should see his online presence, and I should see some contact or info like any data that is on his openID. 

I know that things like Quicksilver and Launchy do this a little but in terms of process optimisation they are still off the mark. They are also not integrated into many systems that people do not really realise they use, especially online services. 

This sort of tool could greatly increase the effectiveness of things like GTD by providing a really 1 to 1 connection with the information that is known by a user and the tasks they want to perform. 

The web is the world and seemingly a large part of the future of computer platforms. If we want this to be the case we should make good tools to interact with it. All the current browsers and most of the browser extensions that exist today do not take good consideration of this fact. I think this feature is one of many that could be used to improve the human interface with the web and the world. 

More on why I think browsers suck later. 

Small Services and Topology for the Same

Today I was trying to work with a few groups of friends on a few different projects at the same time all via the internet. We were doing a few things; research, content building, organisation and  communicating. None of these things by them selves are problematic and especially if they were been done in personal collaboration or by my self then no problems would exist. In this case however, it was very difficult to suitably take part in this system. I found that many of the small tasks I wanted to do could not really be done without a prohibitively large set-up time and more complex things I wanted to do could not be done without more sophisticated web tools. 

For example, I was working away and realised I had an idea for a way to deal with something I was not working on. So I wanted a tool that I could simply write a snippet of text into that would be like a post-it note. Not a big problem, but, I wanted this note to also be connected with the task I was thinking about and annotated so it would be seen by the relevant team members. Now I know there are tools that do this, and services that offer a lot in this respect but I can not afford to sign up for every service just so I can do a few different tasks, and more so, I can not expect all of the people I work with to do the same. So, the answer I think is appropriate is essentially a set of operators on basic communication networks and offer various things like content tagging, targeted messaging, really fast content dropping and so on. I think using as many existing networks as possible is really important, i.e. using email and im instead of creating a new notification methodology, such as the Facebook Inbox, because this approach lets messages go to non members and people who just need to know the content not anything else. 

A different example was when I wanted to create a simple and quick list that included images and some text, it needed to be sortable and searchable but nothing fancy. Now this can be done in many tools like Google Docs or Zoho but all these interfaces are pretty slow and arduous. Also, I wanted to make this list in a way that would make it useful for me later, so potentially I could get a feed out of it or at least use it as a data store for a website. In any case, nothing was really offering me what I wanted so I ended up doing it in a context that had the advantage of local speed, Compendium. Here i could make it both a graphical and context driven list or layout and get rich relationships between elements. Now I know I will have to remake the work later which I think is quite lame. So, I want a tool that lets me have more flexible content manipulation and output settings while also having a high speed interface. An AIR application may be a start and perhaps building it to provide basically a really fixable content bucket that has a hole bunch of import and export tools that are all based on standards. 

I hope open standard interface languages become more popular. I really think they are the answer to a lot of serious problems. But more than that,  I really hope people start to take advantage of them. RSS is a wonderful tool but very few services use it to its full. Similarly, dynamic OPMLs , lists of lists, could be so well used in certain contexts. 

I think almost none of what I have said here makes any sense. I really just want to say. 

1. Service should do one thing perfectly and nothing else. 
2. Everything should be built to work with other things in as many ways as possible.
   
3. The things things work with should be up to the user or the receiver if that role exists. 
4. Standards should even be used for preferences. For instance, how do I like being notified by a tool? The answer to this question should be used when I need to be notified by someone else's service that I have never heard of.
5. Naming and interfacing should be search to activity. (This I will talk about later)
6. Relevance should exist for everything. Like realtime, automated discourse mapping of everything that is relevent.