Zappini

Juli and I are heading down to Portland this evening. I'm pretty excited.

Here's my dailyKos diary entry announcing the event.

Brian Moran has invited me to blog about the event over on WashBlog. Stay tuned.

[Updated 10/7: fixed URL to my diary entry]

I'm running Apple's iTunes 5.0 on Windows. It's both too simple and too complex. Oh, and, yea, the new look blows.

I haven't bought any music online. In fact, I'm still working through my CD collection. Ripping, listening, and rating. What a chore.

Some things that would help: listening history, songs listed in shuffle order, and smarter playlists.

I work while I listen. I also get up and move around a lot. (Fidgety.) So I usually have to step backwards to add ratings to recently played songs. That means I have to stop listening to the current song. That pisses me off.

Hitting shuffle should reorder the songs in the current playlist. When iTunes is bouncing around the playlist, I have to hit that flipping tiny "go to" arrow to bring up the current song, just so I can rate it.

Alternately, I should be able to change the rating (and other info) of the currently playing song.

It'd also be nice if the currently played song in the playlist could be visually indicated somehow. The tiny speaker icon doesn't cut it on a 19" monitor showing 1600x1200.

Lastly, the smart playlist behavior is suboptimal. I have one defined to show me 75 recently added songs that haven't been rated yet. If "live updating" is selected, the moment I change the rating, the song stops playing and the list gets updated. Grrr. If "live updating" is not selected, the list never refreshes and there's no manual override. So I end up toggling "live updating" to effect a refresh. Two reasonable usability enhancing options would be a refresh option (on the context menu) and to have the smart playlist refresh whenever iTunes starts up.

Maybe there are better strategies or workarounds. After a bit of experimenting, this is the best I've come up with. I can't be the only person who has to chew through a couple 100 CDs. Apple should definitely be continuously doing usability testing. Honestly, the ongoing lack of refinement for iTunes is pretty disappointing.

In software development, the process for categorizing and prioritorizing defects is typically called "bug triage". As with chocolate chip cookies, everyone has their own very favorite recipe. The linked article, an excerpt from Scott Berkun's book The Art of Project Management, details an "ultrasimple triage" strategy.

I think these types of articles are great. It's always important to look around to see what everyone else is doing.

If Berkun's triage practices are "ultrasimple", then I guess mine should be called something like "drop dead simple". Or maybe "how to stop arguing and get things done".

Please allow me to explain. First, some background.

I was once a QA manager for a while (until we could fill the position). It was kind of fun. I read what I could. Went to seminars and user groups. Where ever I went (including the bus), I asked pretty much everybody "How do you guys handle QA?" (I eventually stopped asking. Discouragement, see. I never got a novel answer.)

What I learned is that pretty much everybody does it wrong. By wrong, I mean both wasting resources and not accomplishing the mission. I think it's the result of starting with the wrong assumptions.

One popular worse-than-bad assumption is the notion that you have to categorize and prioritorize bugs. It sounds good. How do you control something if you don't organize and measure it first? By categorize, I mean adding labels to bug reports like "Bad", "Really Bad", and "Nuisance". But we software types use important words like "Severe", "Fatal", and "Minor". Another category could be determining which portion of the product the bug appears in.

There's just three (show-stopping) problems with this common strategery:

#1 - If you're spending any effort whatsoever trying to figure out what to call stuff (e.g. develop a taxonomy), you're wasting time. That activity is not critical path, so don't do it.

#2 - QA people shouldn't be speculating over what areas of the product a bug comes from. That's why we have reproduction steps. Besides, it's either obvious or not. And when it's not obvious, like when two or modules don't play well together, there's really no point in recording what portions of the product were effected. How would you even decide what to write down?

#3 - The ranking of bugs is just an excuse to have an argument. Everyone wants their pet bug fixed first. Developers resent QA people directing their work. The discussion of semantics is inevitable ("What do you mean by 'really bad'?"). Whenever there's a ranking, there has to be a record. That's make-work. The ranking continuously changes. More make-work. So the record (e.g. a bug database) is always out of date. Cause for even more arguments.

So now I'll describe how we did things.

Each triage meeting was broken into two sections. First, we talked. If people had questions about a bug, this is when it happened. Most of the time, it was brainstorming about how to reproduce bugs kicked up from technical support. This part lasted as long as needed, typically less than 30 minutes.

The second part was voting. This takes just a few minutes. Anybody who cared to participate could show up and vote. Each person gets 3 votes. QA showed up with a printed listing of open bugs. You voted by making a hatch next to a bug you wanted fixed. You could use your votes however you wanted, even voting for the same bug 3 times.

The bugs with the most votes got fixed first. The developers left the room with their new marching orders. QA sometimes made a photocopy if they felt the need to have a "record".

(I'm pretty sure I swiped this voting idea from some sort of risk management exercise. But I can't remember where and haven't found a link.)

This approach has some really interesting psychology going for it. Somehow, this voting strategery short-circuits the arguing and a consensus quickly forms. It's like magic. Perhaps individuals can more easily accept the group decision when the process is completely open and unbiased.

What makes it all work, I think, is trust. It's crucial that the developers fix the bugs in (roughly) the order specified in a timely fashion. (The frequency of triages changes throughout the project cycle.) That signals to triage participants that the voting is honored and that their pet bug will be fixed in due course.

I guess that's about it. Overall, we tried to keep things simple. It seemed to work pretty well.

Silent films are just the best. Buster Keaton is one of my favorites. So it was with great enthusiasm that I saw the double feature Sherlock, Jr. and The Balloonatic. Last time I saw a silent film at the Paramount, it was for a showing of Greta Garbo's early films. Happily, the audience has grown since then.

Warren Etheredge of The Warran Report was the MC. He's good at it. Trader Joe's sponsored and raffled off some stuff. The only negative for the whole evening was Professor Blowhard. He just went on and on. Did you know that movies are actually a series of still images! Really! And another thing...! Fortunately, Warren was able to wrest back the microphone, preventing a mass suici

Dennis James was the organist on Paramount Theater's restored Wurlitzer pipe organ. James was terrific. Probably the best silent film accompanist I've ever heard. His timing was spot on. The music was very lyrical, meshing with the stories perfectly. I'm pretty sure it was James' own arrangements. Overall, the organ soundtrack deeply enriched the experience. My only advice is for Dennis and Warren to practice the raffle schtick a bit next time. People really enjoyed the whole routine; really playing it up would be great fun.

The movies were great, as expected. The audience really got into it. Lots of laughs. And it was great to hear children in the audience squealing with pleasure and excitement.

There's two more Monday shows coming up. Highly recommended.

Date: Tues, 13 Sept 2005
From: Jason Aaron Osgood
To: US Representative Lamar Smith (R)
Subject: Patent bill would make sweeping changes

Rep. Lamar Smith-


I just learned about your Patent Reform Act of 2005. I'm particularly thrilled by the change from first-to-invent to first-to-file. Quoting physicist Murray Gell-Mann, that idea is worse than wrong. If the goal is to stop all innovation and economic growth, I must admit you Republicans are doing a bang up job.

Also, I enjoyed the account of your recent town meeting (post-Katrina). Particularly the bit from the life-long Republican who claimed his taxes were a "social investment" and demanding to know why you guys screwed up so bad.

You're right on one point, though. Not enough elected officials have contact with their constituents. At least you have the decency to face your voters.


Cheers, Jason Aaron Osgood / Seattle WA