Blog: October 2007

Most of these posts were originally posted somewhere else and link to the originals. While this blog is not set up for comments, the original locations generally are, and I welcome comments there. Sorry for the inconvenience.

Near misses in user interfaces

Yesterday the "low tire pressure" light on my car came on. (I've never had a "low tire pressure" light before.) The tire gauge didn't show anything to be concerned about and the light went off again after a couple miles, so I'm not sure what that was about. (It was not the coldest recent day, at least at the time it happened.) But there was a fundamental flaw in the UI: it didn't tell me which tire it suspected, so I had to check all of them. Given that they've obviously got sensors in each tire, how much harder would it have been to transmit that information and add four little dots or something to the light, lighting the ones(s) where problems were detected? (Hey Honda, if you do this, position them intelligently -- don't follow the bad design of the burner knobs on many stoves.)

On my previous car, you move the lever up to turn on the wipers and down for a single pass. On my current car it's the reverse. That's taking some getting used to. Neither is obviously better; I wish the industry would just choose one.

Another in the "it's not just about you, mister designer" class: every microwave oven I've ever used has a numeric keypad, with "start" and "stop" buttons to either side of the "0". On the microwave at home, "start" is on the right. At work, it's on the left. As a result I get this wrong about one time in five. (It's not as if I -- or most users, I suspect -- actually read the button; we use positional memory, which works for numbers and fails for start/stop.) People change microwaves more often than they change cars, I suspect, so it would be nice if the industry would settle on a standard. Either one would be fine if it were predictable.

Why we need a configuration manager

The original entry is locked, but it's been long enough that this post can be public now.

I mentioned last week that I'm doing a somewhat-messy code merge at work. Here, let me be more explicit:

scary branching diagram

This is a slightly-simplified view of the code branches my project cares about. A dotted arrow represents a partial merge (individual features, small or large). A solid arrow represents a complete merge (everything on the source branch gets integrated into the destination branch). Arrows with the same color represent the same features, perhaps at different points in time (updates). Arrows skip over any intermediate branches; for example, the blue and purple arrows go straight from the product branches to my branch. Those are completely unsupported by the product team (properly so); we call it "taking crack". Red dots are releases. (I only showed the releases that directly affect me.)

See those brown arrows? We're doing the one pointed at my branch now. It touches, oh, several thousand files. We did things like move foundation classes to different packages and divide some up into interfaces and implementations. (Fortunately, I have merge help from teammates.) The product tech lead and QA lead are doing the other brown arrow, and that branch also has to get to the current product branch. (They took that product branch too early; the plan was to get the refactoring back to main first.)

Big integrations are really done on integration branches first, which I mostly haven't shown. I've included some speculation about how my project and the vertical domain's branch get back to main. Coming back to main requires a fair bit of integration and test, of course.

I've been doing most of the merges onto my project's branch. I'd be thrilled beyond words if we succeeded in hiring a configuration manager to do that integration from main between my two releases. :-)

Techies and management

Someone I follow linked to an article about why it's hard to keep senior developers doing development. The author asserts that no self-respecting developer started his career thinking "I want to be a manager!", but the natural progression is toward technical leadership, management, or both. This progression is captured, somewhat cynically, as the Peter Principle.

The author of the blog article calls on developers to resist the urge. Remember that you love coding, he says. But I think there are good reasons to want to move in this direction, reasons that probably don't become apparent until you've been out there a while, and which I'm seeing now that I'm in that position.

I think I was a pretty darn good individual contributor when that was the main thing I did professionally. Because I work with a cool group of people in a business sector that already tends toward this kind of flexibility, I was generally able to get a hearing with the leaders -- product managers, tech leads, executives, et al -- because I'd proven that I had some clues. That means it was possible to have influence beyond my own little box. That's important to me. If I'm going to spend 40+ hours a week on something, I want to have some ability to make it matter. If I don't care, I disengage, and that's no good for job satisfaction or employer satisfaction. I need to believe that I'm making good stuff and meeting professional quality standards. Nearly a year ago, when the possibility of my current project first arose, I told our CEO-equivalent that if we're going to do this right I want to be deeply involved, and if we're not going to do it right I want nothing to do with it. He made me tech lead.

I've been a tech lead before, but this is my largest and most complex project so far, with potential to get much bigger, and it's been pretty nifty. I'm providing a lot of the technical vision, conducting design and code reviews, mentoring newer people, learning to be agile and adaptive (new requirements, discovering bodies in the code base, bugs, etc), and, in many ways, setting the agenda. I'm not the arbiter of requirements nor long-term strategy (and I don't have customer contact), but I have opinions there and a chance to get them heard seriously. Because I've been at the company a while, I'm seeing seeds I planted (as tech lead of a much smaller effort) years ago finally starting to take root -- this project is a logical follow-on, and in the intervening time I've gained some credibility. I think I can push out some good stuff beyond the confines of my own projects, and this role gives me better traction to do that.

I make sure I keep some individual-contributor work for myself, because I don't want my skills to rot and because I want to produce something beyond carbon dioxide and meeting notes. I also hold by the philosophy that, as much as possible, I shouldn't ask people on my team to do stuff I'm not willing to do myself. I might not be able to do it due to time constraints or (lack of) specialization, but I should strive to be able to, as much as practical. First off, it avoids that peons-versus-leaders chasm where the scut work falls to the so-called peons (today I started doing a particularly ugly merge, for instance). There shouldn't be peons; I want the people on my team to feel they have more power than that. Second, it improves overall adaptability if someone gets sick or hit by a bus or lured away by Google.

I'm also a manager (in the HR sense); I currently have three direct reports, two of whom I got this year. I used to worry that I would be bad at this, but so far that's working out ok. Part of that is excellent people. But I also find that I enjoy the mentor role. (And it's a small-enough group that management qua management doesn't consume a lot of time, except during performance-review season.)

All of this gives me a chance to increase my impact. I think I'm good at what I do and that I have a decent clue about what's good for our business unit on a strategic level, so this seems like a win to me. But it's not without risk; most significantly, I think, it can be hard to go back to the old role. I like this now, and I like where I think it leads for the next few years; beyond that is pretty hard to imagine.

Granted, I was never the type of senior developer the article is targetted at, but there are enough similarities with the sometimes-twisted path my career has taken to resonate anyway.

Interviewed by Tigerbright

1. Do you have a favorite fruit? Why that one particularly?

It depends on context, I guess. I really like the flavor of mangos, but have thus far been unsuccessful in doing anything useful with a fresh one. (I'm not convinced I know how to pick out an optimally-ripe one.) A little less exotically, I also really like the flavor of peaches. For just plain eating raw, though, I'll usually grab either apples or bananas.

2. What's the first SF/F book you remember reading?

That would be either A Wrinkle in Time or The Phantom Tollbooth. I read both several times and can no longer recall which came first.

3. How did you end up learning to play the hammered dulcimer (the first thing I noticed about you)?

I learned about the dulcimer via performers like John McCutcheon, Malcolm Daeglish, and Maggie Sansome. (There were others, but those are the ones who stood out.) I was a recent college grad who thought it was a really cool instrument, but the going price was 1.5 months' rent and I wasn't that dedicated. Then one day I was wandering the Three Rivers Arts Festival with friends, and we stopped at the booth of a local instrument-maker. I played with the dulcimer that was on display but decided against; one of my friends was also casually interested. And then the guy manning the booth said that, actually, they were getting out of the hammer-dulcimer business (mountain dulcimers and harps were more popular among their customers), and they would sell their last two for the price of one. That was just what I needed, so we bought them.

Mind, it was a low-quality instrument, and in retrospect I'm surprised I kept it in tune enough to learn on. But it was enough to get me started, and a few years later I went to House of Musical Traditions (near Baltimore, or maybe DC -- even now the cities are somewhat mingled in my geographical hind-brain) and bought a better one made by Dusty Strings. It was a few months' rent, but at that point I didn't care.

4. What's your favorite song to perform, with or without backup?

Wow. Just one? Ok, of the songs I have actually performed (as opposed to the ones I sing in the car or shower), I think it would be "Some Kind of Hero" by Misty Lackey and Leslie Fish. There are other songs as well-written, with as much emotion and power, and that fit my voice well, but for reasons I can't quite identify, that one still tends to come out on top. I'm not sure why.

Pegged him

SCA friends, both serious researchers, just got back from their month-long honeymoon in Scandinavia. One of them works at my company, so yesterday he was chatting with folks about the trip. When I joined the conversation, another coworker said to me "ask him how many pictures he took".

Wait, I said, let me try to guess first. I asked the coworker how many museums he went to, and he said "one or sometimes two a day", but they lost a couple days to travel. So, I said, average of one museum a day for 30 days? He said that was about right, and volunteered that almost all of them allowed photography. We had already established that he was using a digital camera, not film.

Ok, I said. I'm going to estimate that you took between 250 and 500 pictures per museum, probably closer to the former. (The other coworker's eyes get big.) So, I continued, if I take the conservative number for now, that's 7500. I'm guessing you took a few picturs of scenery and stuff but not a lot (other coworker looks startled), and I did start with a conservative estimate. So, let's say between 9000 and 10,000 pictures total (other coworker's jaw drops). How'd I do?

He's not sure, but he knows it's at least 9000 and he thinks not as many as 10,000. :-) The other coworker asked "how did you do that?!". I said "you just have to know him". (Besides, I showed my work.)

Random bits

I've noticed that when there is a great feline tussle in my house that leaves piles of hair around, the vast majority of the time the hair belongs to Baldur. I see several possibilities: (1) his greater surface area makes him more likely to be hit; (2) his hair just doesn't stay attached as well as the other cats'; (3) he gets picked on a lot (he's the biggest cat BTW); or (4) he has developed the "eject hair" escape technique. Hmm.

I missed the first episode of the new TV show "Pushing Daisies" but caught the second. Wacky! Surreal! Fun! The narration as commentary is a nice touch. Yeah, that it's written by the person who did "Wonderfalls" shows; I hope "Pushing Daisies" fares better. ("Wonderfalls" was great for about 8 or 9 episodes, then sucked for a couple more, and was then pulled after 13.) I'm also watching "Journeyman", about which I'm undecided.

We drove through the rockslide zone of Route 28 on the way to visit my parents today. No rockslides were in progress at the time, and it looked like last week's had been completely cleared. The news had said inbound lanes would be completely closed for the weekend, but we saw continuous traffic while we were driving outbound so we didn't look for an alternate path home. It turned out that one lane was open. That was fine for a Sunday, but I'll bet it sucks for commuters right now. That said, rockslides suck more.

Two Shabbatot ago a first-time (in our minyan) Israeli torah reader asked me to be his checker. I expressed concern that I wouldn't be able to keep up; he said he reads holy texts slowly. His "slow" was too fast for me. Then this past Shabbat a different reader asked me to check for him and I figured this wouldn't be a problem; I had just a bit of trouble keeping up. Both times I was checking from the new Plaut (oodles better than the old Plaut), and using a magnifying glass to be safe. I conclude that my problem is Plaut + magnifier, not necessarily me, and I should only check when I can do it from larger Hebrew text such as what Trope Trainer produces. (I'm not the only torah reader in our group who uses that software, and in fact I have been handed TT output to check from at times.)

Without saying anything about the merits of Al Gore's work, I do admit to being puzzled by how this is a Peace Prize issue. Of course, in political processes all bets of rationality are off, but still... isn't there a more appropriate category in which to consider his work?

I heard a cute story recently: One night at dinner the seven-year-old girl asks her parents "where did I come from?" Oh crap, the parents each think; we thought we had a few more years before we'd have to deal with this. They exchange glances and then fumble through a discussion of birds, bees, and what happens "when mommies and daddies love each other very much". The girl says "oh" and everyone sits in silence for a few minutes. Then she continues, "my friend Becky comes from Cleveland".

"Private" vs community bar-mitzvah services

Judging by the traffic on relevant mailing lists, lots of Reform congregations have the problem of families expecting to "own" the service at which their kid is bar or bat mitzvah. ("Own" means the kid does most of the service, family members get all or nearly all of the honors, the parents stand up and kvell about the kid for several minutes, and so on.) The topic came up again this week, with someone asserting that we have to make kids feel welcome and this "cannot be done" by a service not owned by the family.

I find myself wanting to write about this from time to time, so I'm recording my response to that message:

If we are to embrace and welcome the students as they attain their next level into adulthood and make them feel wanted, we must support them in their studies and praise them for their accomplishments. This cannot be done by 'community owned' services.

It can be and is done. I have been to services in several congregations where a bar or bat mitzvah was an integrated part of a community service, with the family neither "owning" the service nor being sidelined. It can work. I have only seen this once in a Reform congregation (Holy Blossom in Toronto); it is the norm in Orthodox and many Conservative congregations, and it worked beautifully the one time I saw it at Shir Chadash (traditional egalitarian) in Jerusalem.

The problem is the following vicuious cycle: families in the "it's all about me" generations in America demand ownership and won't participate in a community service, they get this, as a result the community doesn't come to the "private" service, and so more families feel justified in this expectation ("they don't come anyway"). This is made worse by the fact that most of our members don't see our congregations as important communities; I see much more of a "consumer" attitude. If you're part of a community then of course you want to celebrate your milestones together; if you see the synagogue as the place where you buy services such as a bar mitzvah, you're less likely to be interested in what the community wants or needs. I'm not pointing fingers; this is just how it looks from here in the pew, from someone who's there pretty much every week and sees who does and doesn't come regularly.

I don't know how you stomp out the "private bar mitzvah" once it's present; congregations that have never let it take root do not seem to have a problem with b'nei mitzvah feeling, and being, welcomed into the community. And I sure don't know how we fix the broader problem of community engagement.

From comments (archive):

Part of the reason it's generally not been a problem in the (Orthodox) congregations I've seen is that the kid/family is part of the congregation before, during, and after that weekend

Bingo. The family is part of the community, and naturally the community wants to celebrate milestones for its members. Alas, many families join synagogues just to get the bar mitzvah, come rarely, and disappear right after; the sorts of folks with this attitude are going to seek out congregations that have less of a "full community" meme to begin with, so (1) that's more likely to be larger liberal congregations and (2) by doing so they help increase that perception for the next such family, and now the cycle is off and running. :-(

My syngagogue certainly has communities within it, most notably (to me) the informal Shabbat morning minyan. We're there for Shabbat, for study and prayer. This group celebrates its members' milestones, though we haven't yet had a bar or bat mitzvah because not many kids come. But we're a 40-50-member community within a synagogue with over 800 households, so most of our members never see this. (It's not for lack of outreach; they're just not interested.)


I think part of it depends on whether the Bar/Bat Mitzvah is at a regularly scheduled service or not. If not, there is more the perception that this is a private service.

Indeed. We have our 10:30 Shabbat morning service only if there is a bar or bat mitzvah (which, to be fair, is about 75% of the time). We have the informal, partly-lay-led service at 9:00 every week. So the 10:30 service tends to be the family show. However, we also read torah Friday nights at our regular community service and, a few times a year, there will be a bar or bat mitzvah at that service. And the families behave the same way as they do at the 10:30 morning service, except we only do 3 aliyot and not 7 and the rabbi encourages them to keep the parental kvell short (which some blow off; I've seen kvells that run 10 minutes, which is at least 8 minutes too long).

It seems to me that that Friday service is our opportunity to change things. Some families view a Friday-night bar mitzvah as a booby prize (though we never assign it, so they asked). We need to change that perception; a Friday-night bar mitzvah, when you have the whole community there, should be an honor. We should take the families that "get it", the ones who understand that this is a community and not Egoboos R Us, and ask them to celebrate their b'nei mitzvah on Friday nights with the community to model the desired behavior. Maybe, over the course of a decade or two, we could effect a change in attitudes.

Eye tests

I like the technician who did some routine tests for me this morning. She talks and shares reports. :-)

They have a new machine for visual field scans, so instead of taking 10-15 minutes (feels like 30) per eye, it now takes two minutes per to map the field of vision. The interface is still the same: stare at the cross-hairs and push the button when you see a flash of light anywhere. Flashes vary in position and brightness. I saw diagrams of the results showing the location of my blind spots (apparently we all have them). The machine still has the unfortunate characteristic of making noise when lights go off, which (1) makes it obvious I'm missing things and (2) enables gaming the system for those so inclined. (Gaming is obviously bad if you actually care about accurate results, but if you're a little unscrupulous and just need to pass this test to get your pilot's license, say, there'd be incentive.) Maybe they're actually clever and some of those sounds correspond to no-ops; that would be good.

Normal cornea thickness is 5.5. ("Units?" "Um, I don't know, but I can look it up for you." "That's ok; I'll ask the internet.") Both of mine are thicker than normal (6.26 right, 6.4 left). This is better than being too thin, and also means that my pressure isn't really as high as the glaucoma test says it is. I don't know if this is well-enough understood to be able to compute my actual pressure given a reading and the thickness data.

Learned in passing: the magic number for worrying about high pressure is 20. I always test in the 18-22 range, by the way. (That's with treatment; don't know what it would be without. Not inclined to find out.) So if I'm hovering at the magic line and my real pressure is probably a point or two lower, that suggests that the treatment is calibrated right.

She also took pictures of my retinas and let me see them. We had a lot of trouble with this; apparently my pupils were clenching their little fists and whining "we don't want to be dilated!". A double dose of the drops was enough to get a good picture of my right eye, but the one of the left was still kind of dark. If it's not clear enough, they should be able to redo it when I'm there for my next exam.

Daf bit: Ketuvot 40

(Preface: instead of doing another year of parsha bits I decided a few weeks ago to do teachings from the talmud, following the daf yomi cycle (even though I'll be doing this weekly rather than daily).)

In discussing laws related to marriage, the talmud raises the question of the precedence of positive and negative commandments. Rav Kahana says the positive commandment always supersedes the negative one, but Rabbi Zevid of Nehardea says this is only true when performing the negative commandment would make it impossible to perform the positive one. He gives the following example: it is forbidden to remove tzara'at ("leprosy", but not really) by surgery, but should it occur in a certai place, we are nonetheless required to perform the positive commandment of circumcission. (40a)

I didn't import most of these "daf bits" to this blog. You can see all of them on Dreamwidth by clicking on the tag. This was the first one.

Simchat Torah

We got a huge crowd for the evening Simchat Torah service (Wednesday night). There were lots of kids; someone had organized a gathering before services where the kids could make paper flags to parade around with, and I suspect that made a difference. (We always get lots of kids, but the "lots" seemed bigger this year.)

Lots of kids means lots of noise, especially when you factor in today's trend toward permissive/oblivious parenting, and that continued through the first two aliyot of the torah reading (read by our two rabbis). As I went up to read the third and fourth aliyot (from B'reishit), I mentally braced myself for the distraction.

I should mention that the previous aliyot were read, not chanted. I began chanting -- and everyone shut up, just like that. Wow! I don't think they were just being polite to the layperson; I must conclude that it was the unexpected music that did it. I won't complain. :-)

I got compliments that night (and the next morning when I repeated it), whic is not uncommon (it's polite, after all). The person who stopped me on the street to compliment me was a little unusual, and the rabbinic student who told me Shabbat evening that he was especially pleased with the way I'd chanted one phrase made me happy. It's nice to know that those little details matter to people other than me. :-) (This was, in fact, a passage I'd worked on to get the expressiveness right.)

The hakafot (dancing, or at least parading in a somewhat-perky manner) were more lively this year than in the past. We've been attracting more Israeli members (I suspect because of our Israeli rabbi); maybe that's part of it, as they bring in the enthusaism that (I'm told) characterizes this holiday there. Some of it is coming from the rabbis and cantorial soloist, for sure. I think we've got a fair number of congregants who will follow a lead to cut loose a little but won't lead, which isn't surprising.

Thursday morning during one of the hakafot I had a sefer torah that had silver crowns on it (not all of ours do), and the crowns had bells on them. So I could produce sound effects in time to the music. :-) (The scroll was too heavy for me to safely hold over my head, alas, especially as it was wound to one end. At other times I had smaller ones that I could do that with.)

Learned in passing: we are one of three (Reform?) congregations in the US that has a complete set of Nevi'im (Prophets) and Ketuvim (Writings) on scrolls. (I do not know how this came to be.) We pretty much only use these on Simchat Torah and Kol Nidrei, alas; most bar-mitzvah students want to read their haftarot with the permitted vowels, which you can't really blame them for. (While we read torah on Friday nights, we do not read haftarah then.)