Another option for how to link up 'thoughts' is to use something similar to suggested stories on a news site. These suggested thoughts would be related to the current thought using an configurable algorithm. This would reduce the amount of friction when adding thoughts and encourage me to do it more.

Rather than complicating the UI the 'suggested thoughts', which from now on I will call 'related thoughts' because the verb suggest doesn't fit in this context, can fit in to the existing stream of thoughts, above and below the current thought depending on the creation date. This is where 'infinite scroll' would be very useful.

To improve the utility of the Doing widget I should start using it more for personal tasks. I think the task description, which for work items I currently don't read very often, would be more useful and I would engage with the actual content on the widget rather than the colours and the notifications.

To encourage we to use it for personal tasks I think the following improvements might help.

1. Make it easier to enter thoughts. This can be done by
   1. Making 'Add Card' work offline
   2. Add new cards to 'This Week' or 'Today' depending on which one is current in view.
2. Notifying the user when 'This Week' and 'Today' lists are empty in certain time windows.

I need to tweak the 'More Like This' configuration so there is a higher threshold for the related thoughts.

It would be nice to be able to reference previous thoughts when writing new ones.

There are two approaches that initially come to mind to tackle the presentation of referenceable thoughts.

1. Provide links to the referenced thoughts and referencing thoughts along with the current thought.
2. Present all referenced thoughts along with the current thought and make the referenced thoughts collapsable.

Thoughts are not easily categorized, however as it might be helpful in deciding (to whatever extent we do decide) which thoughts to act upon and how, I will define three broad groups which thoughts could be sorted into.

1. Thoughts on past experience, something like how your memory is realized.
2. Thoughts about how potential events could play out, be it due to your intervention or not (like a plan or prediction).
3. 'What if' thoughts, could be regret or something less emotionally charged.

Thoughts from all groups appear to invoke emotion, so a memory of a nice event or promising plan could invoke a feeling of happiness and/or satisfaction but emotion could also be genesis of the thoughts. Although the genesis of a thought never seems clear to me.

Writing thoughts where it is assumed I am the consumer seems to be more satisfying.

I recently implemented a 'context service' for the PocketLab web app which has proven to be very useful for multiple reasons.

The service, which is part of an angular web app although assumably could be used in most front end web frameworks

1. Listens on URL changes
2. Constructs a context from the new URL
3. Publishes the context through an Observable.

In the web app I am working on the context consists of the object currently being viewed and the hierarchy it exists in as well as some more broader information such as login status and PocketLab device in use.

Amongst other things this is useful for providing information to UI components that help the user orientate themselves within the app and for sending a lot of rich data to Google Tag Manager.

A couple of thoughts regarding the PocketLab UI.

1. The collection of data across different data sources should be more closely integrated so the user can easily select between different graphs from different sources on the same screen. This would be similar to how post collection data is reviewed across sources.
2. The revision of data (post collection) should be more distinct from the collection of the data.

Some processes I perform during the day that I might be able to improve on are:

1. Making coffee
2. Exercising
3. Reading
4. Making dinner
5. Programming
6. Managing projects
7. Betting

Improvements could entail

1. Making the process faster to perform
2. Making the result of the process better
3. Achieving more enjoyment from the process

For each process one type of improvement may be more desirable than another. For example, I don't want to speed up my coffee making as I use it as a chance to take a break from looking at the computer screen but I do want to speed up project management as it would give me more time to do something I enjoy.

Although I believe a more liberated engineer (free from excess social, bureaucratic and traditional processes) generally makes for a more for-filled and ultimately more productive engineer it can also be helpful to use self imposed mechanisms to guide behavior where human tendencies can lead us down unproductive paths.

An example of this is a widget I created for android which connects with my Trello boards, sits on my home screen and has the following features:

1. A Doing section, (scrollable) Today/This Week section and Clocked off section which track corresponding lists on all Trello boards and present the cards present in these lists.
2. Ability to move the cards between lists as well as into the 'Todo' and 'Done' lists which are present in Trello but not presented in the widget.
3. Triggers notifications and alarms when amongst other events the following happens
   1. A card sits in 'Today' after the day is complete
   2. A card sits in 'This Week' after the week is complete
   3. A card sits in 'Doing' outside work hours
   4. There is no card in 'Doing' during work hours

The alarms and notifications have a pavlovian effect and can be made to quite unpleasant by adjusting vibration length and sound thus increasing the effectiveness.