In DestroyTwitter 2.0, the Account canvas gets a major overhaul. As much as it was nice to change your name, bio, and location in 1.x, I found most people only used it once, if at all. I also discovered that more users were interested in seeing their own profile and viewing it in the People canvas made it a pain, constantly needing to reload.

Because of these issues, the Account canvas in 2.0 extends the People canvas. It displays the active user’s info, including API status, in semi-realtime. This means any time you login, tweet, send a DM, receive a DM, follow someone or unfollow someone, this info is updated without using an additional API call. The API status is actually updated any time DestroyTwitter receives any data from the Twitter API without using an additional call. Also, when you click a user’s icon or username in a tweet, if it’s the active account, DestroyTwitter will know to go to the Account canvas rather than the People canvas.

This screenshot does lack a few buttons I plan to add, such as logout, but it’s mostly all done. I’m really excited about these fine-tunings and improvements over 1.x. Do realize I’m working every night to get this out the door. It’s coming along, but will endure a solid private testing period. I’ll make an announcement soon regarding how to sign up. In the meantime, hold tight.

The biggest feature in DestroyTwitter 2.0 is multiple accounts. Because of this, a serious amount of planning went into memory optimization. I didn’t want usage to double with each account added, so I adopted a few approaches that worked beyond my wildest expectations. One technique I used is improved item rendering. A few weeks back, I wrote a post about scrolling methods. The current one used in DestroyTwitter 1.7.2 uses preallocation and masking. It’s great for scrolling speed, but awful for instantiation speed and memory usage.

In 2.0, I adopted what I call fixed-height smooth recyclable scrolling. Instead of 200 list items for 200 tweets, I have only the visible items and a data list of 200 tweets. So far, it’s proving its worth. In recent testing, I’ve seen a 20mb memory reduction when using three accounts simultaneously in 2.0 compared to one account in 1.7.2. Results like this make me a happy developer.

Here’s a first glimpse at DestroyTwitter 2.0. Relating to yesterday’s post regarding preferences, I decided to segue into the Preferences canvas. In the current release, this canvas is a bit long-winded and claustrophobic. To combat these issues, I adopted an accordion layout where each set of preferences is expandable and collapsible. Not only does this clean up things a bit, but it also allows me to reuse components, which reduces memory. And finding new ways to reduce memory is like Christmas come early.

With multiple accounts in DestroyTwitter 2.0, there are a number of things to consider—one being preferences. In the current release, preferences are global across all accounts. In 2.0, however, some preferences are specific to each account. Above is my plan for divvying them up. Most rest under application, constant across all accounts, but there are a few that might differ from account to account.

Settings like refresh intervals are a given since some accounts require more frequent updates than others while some have canvases that don’t even need to update at all. The ability to disable Quick Friend Lookup when typing “@” is a new preference that is specific to each account. For instance, I want this enabled for my @destroytoday account since I frequently tweet to friends. On the reverse, I want it disabled for @destroytwitter since I only tweet to someone who asks a question or submits feedback.

Over the next few weeks, I’ll write a post previewing the upcoming release. I’m hoping to get it out this month, for private testing at the very least. Check back often for insight into what to expect.