Things that were not immediately obvious to me

Today I’m going to present a workaround for a slightly obscure but still annoying problem on the iPhone. Let me begin by quoting myself:

If you create a view controller, put a UITextView inside a table cell in that controller’s view, make that UITextView the first responder, and then push the new controller onto a Navigation stack, the keyboard does not scroll in from the right; it pops into view immediately. That’s ugly …

It turns out that, with a little hack, you can make the keyboard slide in properly.

Continue reading →

A word about the BEGINSWITH operator of NSPredicate: “Caution”! I was using this operator to drive iterative prompting while developing my Full-Text Search (FTS) demo, and found its performance unsatisfactory.

Continue reading →

Heretofore, I’ve been working with an “OR” implementation of Full-Text Search (FTS). This sort of implementation takes one or more keywords, and returns all the objects which match one or more of them. I’d been avoiding an “AND” implementation (which only returns objects which contain all the specified keywords) because I suspected that the Core Data/SQLite combination would make a hash of the necessary query plans, but this turns out not to be the case. So today: “AND” style FTS!

Continue reading →

Today I’m going to post minor revisions to some earlier Full-Text Search (FTS) code. These changes address index construction.

Continue reading →

I want to return to yesterday’s topic, and look at further optimizing the Full-Text Search (FTS) query. Although we’ve addressed the gross inefficiencies caused by suboptimal query plans, there is still some fat to trim; it turns out that the sorting of results can be a significant factor in query performance.

Continue reading →

Yesterday, I covered some of the more interesting parts of a prototype implementation of Full-Text Search (FTS) on the iPhone; today I want to start to address one of the performance problems in that prototype. To begin with, it takes a ridiculously long time to execute a search on the index. This turns out to have a simple remedy.

Continue reading →

Today we’re back to implementing Full-Text Search (FTS) on the iPhone. On Monday I presented an FTS prototype written in Python, and since then I’ve been working on porting it to Objective-C, targeting Cocoa and iPhone OS. I’ve implemented a working prototype on that platform – albeit one that runs very slowly on the device – and today I’d like to present some of the more interesting parts of this work-in-progress.

Continue reading →