A common need in FRP is pulling in another stream inside another stream and combining the results. In some frameworks, this is super easy.
I’ve been working on a project that’ll be used throughout my client’s entire organization. That means the code needs to be extraordinarily understandable. Here’s a few observations (ok, one anti-pattern and one recipe)—more may come in the future: Continue reading
I friend of mine recently asked me for career advice and that my wheels spinning…what are the challenges to working in the computer industry? and how can you be successful in the long term (+10 years)?
- a definition of a nueron (so that we’re on the same page)
- an implementation of the backpropagation algorithm (batch mode)
- links to neural networks and the backpropagation algorithm
As usual, there is code! https://github.com/frankandrobot/mustached-octo-neuralnetwork
(Oh and as of 12/8/2013, my notebook is visually ugly. Will style it later.)
Ever since I wrote the new ContentProvider framework, Reminderer had this annoying bug where it wouldn’t auto-refresh the open tasks list whenever you added a new task. You added a task but the darn task wouldn’t show on the list until you rotated the phone.
The first approach was to force reload the Loader every time the fragment gets focus. While this hack works, its just that—a hack. This approach quickly becomes inefficient once the task list gets unduly large (as in hundreds of tasks). The second approach was to use a ContentObserver to listen for changes in the database. Fortunately, for some reason, as I’ll explain shortly, the content observer wouldn’t listen to changes in the table.
I’ve spent the past week fixing some persistent and annoying bugs. More on those later. Being a fan of the KISS principle, I think its better to polish what I have so far before adding more features rather than… adding a gazillion features that half work. So you can follow the latest progress in the dogFood branch.
The immediate goal is to make the “open tasks” activity more polished: