Last night we deployed a pretty substantial update to Tempo, our time tracker. Many of the changes are unseen, nuanced, and/or performance-related, so I want to give our customers an update on what we’ve done. It looks like a short list, but it was quite a bit of work. Details follow for some of the more significant changes.
This was a huge update. In order to support future planned features we needed to overhaul our handling of time localization. There were a couple of hiccups last night during the deploy related to this, but we’ve got it worked out now. Tempo does browser-based detection of your time zone for new users, setting your time zone preference. On the Account screen you can now change this preference, should you so desire.
As mentioned last week, we’ve done a bit of work to make sure that our remote command message handlers are rock solid in the face of weird input via email, and outages on Twitter. The remote command handlers are Looper daemons that check to see if you’ve sent in time entries or timers by email or Twitter.
Also mentioned here previously, the batch tagging interface received a serious speed bump. This was a bit tricky, because we’ve been caching tags as a concatenated string on the entries for performance. Nothing a little PL/PgSQL can’t fix.
It’s no secret that Tempo is a RubyOnRails application. It started it out on 1.2 and we’ve kept upgrading with each update to the framework; we’re now on the 2.3 series. This brings significant performance and flexibility improvements. It may not seem like a big deal, but the amount of work that goes into making these upgrades a success is well worth the sweat considering the capabilities it opens up for Tempo in the future.
That’s that for this update, but there’s a lot more coming around the corner. We have a new feature that’s almost ready, and we’re working with nGen Works to develop a new graphic design and improve the user experience dramatically.