To determine the extent of the changes that will be required to have the call log as the opening view for the Dialer, we should prototype it. We must be certain to lazy-load code that was previously required immediately on load and no longer is. Once this is ready, we should gather the following data: 1) Load time according to the same measurements used by Datazilla. [1] 2) Render time for a single log item -- we can do this without the prototype. 3) Render time for some arbitrarily large number of log items, say 500. Here are our objectives: Primary - Stay under the 1000 ms load time requirement, as measured by Datazilla. We are currently taking about 550ms to load. [2] Secondary - Stay as close as possible to our current load time. This is somewhat idealistic as it is unlikely that we'll be able to stay at or near the current load time. Once we have all the data, we can then make a decision as to what the best way to tackle this problem is. The list of "depends on" for this bug are each potential ways that we can improve load times. We can mix and match these to get the desired result, or come up with new ways if none of these address the problem. [1] https://wiki.mozilla.org/B2G/QA/2014-11-07_Performance_Acceptance#Startup_-.3E_Visually_Complete [2] https://wiki.mozilla.org/B2G/QA/2014-11-07_Performance_Acceptance#Dialer

Whoops, these should really be blocked by this bug instead. We can make one or more of these blockers of bug 1097757 once we've determined which ones are needed.

We might benefit from doing more lazyloading outside the region of the call log.