Google Correlate

My friend Rob Hatton pointed me to Google Correlate, a new service that lets you explore data trends in a few different ways. The theory is that the history of searches should expose both some interesting trends and correlations, often faster than than the way ‘official’ approaches work.

On a lark I typed in SQLSaturday. It found no matches but recommended SQL Saturday, so I tried that, and got the chart below. I’ll have to admit that I didn’t expect a correlation to Dave’s Killer Bread! I’ve amused myself trying to guess reasons:

  • We have a speaker who loves DKB and is evangelizing (though I had not heard about it until now)
  • Spouses are waiting for the weekend of SQLSaturday to buy the DKB so they can have it all to themselves
  • Dave is a SQL guy
  • Dave is an incredibly savvy marketer that sees the grass roots success of SQLSaturday and has found a way to associate his brand with SQLSat (and now I’m helping!)

Beyond my mild amusement it’s something I wish I had more time to spend working on. They have a really nice data set and are giving us some tools to work with it without working about technical trivia. It may not answer every question, but even if it helps someone discover the correlation and then apply more sophisticated tools it’s a win. It will be interesting to see.


Energy Efficient Offices

I work with a client that recently moved into a newly renovated space in a smaller office building. The renovation was nicely done, and as part of that they put in some effort to make it energy efficient. Compact fluorescent bulbs in the hallways, some new windows for natural light, and motion switches for overhead lights in most rooms. It’s a nice space.

I’ve had the chance to visit the office a few times now, and it’s interesting to think about how it works. I think the first thing I noticed is even though it was a complete renovation, there are no hands free faucets in the bathroom. That makes me wonder – does it cost too much to put them in or operate them compared to what it would save in water? The windows aren’t tinted either, does the tint lower the natural light in the office requiring most lights to be on, or was it too expensive to apply, or just not something they thought of (in Florida?).

There’s a humorous side too. My friend tells me that if you, ah, spend too long in the bathroom the timer turns the light out, and of course at that point you’re hidden from the motion sensor!

Attention Speakers–Sign Up for SQLSaturday #85 in Orlando!

Our call for speakers for #85 will be open about another 10 days, and then Bradley Ball can jump into the fun and chaos setting up the schedule. As always we’re going to try to give everyone one presentation and we’re friendly to first time speakers too!

Reasons to speak at our event:

  • We’re the only SQLSaturday that weekend (no one wants to compete with us….hmm, probably not, but sounds good)
  • Speakers will be getting an interesting gift. Last year it was a very nice coffee cup. This year, something we don’t think you’ve received yet!
  • Weather in Orlando in September is gorgeous. Send the family to Disney World while you’re here!
  • The speaker party. Want first class networking, we’ve got the location for it, perfect for long talks and dinner
  • The fun of emailing Brad with schedule changes and suggestions (don’t under estimate this one)

Seriously though, we have a lot of fun, low key, good logistics, lots of low cost hotel and travel options if you decide to visit. I’m looking forward to seeing a lot of old friends at the speaker party and making some new ones. Hope you can make it.

PASS Update #58 (Elections)

Last year we had some problems with our annual election, enough problems that clearly we needed to do a deeper look at our ‘process’ to see what was working and what wasn’t. We ended up fielding an Election Review Committee (ERC) that had it’s first meeting at the Summit last year and continued meeting by phone up through May of this year. Bill Graziano has posted a good summary of the results of that effort on his blog.

I served on the committee and it was close to grueling. It’s a challenge to meet for an hour by phone every week or every other week, keep track of where you left off, and put some time into thinking about things that aren’t clear. It was a good group, diverse opinions, willing to listen, willing to think, and most members made most of the calls, not bad at all for a six month effort. Some notes about the process:

  • The biggest mistake we made was not having some from HQ or a temp on each call to take notes. It’s important to capture the notes and get them out in a timely manner, we missed this a couple times and it hurt. Especially over a long period having good notes (minutes) is important.
  • This wasn’t something that could be in a weekend. I think a good weekend session early on would have been helpful, but some of the stuff we discussed needed time to bake. The same for phone calls. I think maybe 1.5 hours might leave a little more wiggle room, but longer than that would be unproductive. Reducing the scope would have been difficult,it’s a holistic process,you have to think about how x affects or mitigates y.
  • We didn’t get a lot of public feedback on the stuff we posted. We had a diverse group, but part of the goal of any group is to drive toward a shared view. There’s no guarantee that you’re down in the weeds. Maybe no feedback indicates we’re on track, maybe it indicates a lack of interest, and definitely we could have made it more visible. One thing I’d like to try for stuff like this is to set up a mailing list where members opt-in and get mailed occasional surveys to provide feedback on various posts.
  • Joe Webb did a nice job steering. It was a good group to work with, actually very good, but with any phone meeting you have to work to stay on track and Joe helped with that.

In the end I think we have some decent recommendations. I went into it wanting a few things; a write-in that would bypass the nomcom, a clear definition on values we were looking for, and a way to make the process of picking the final candidates less subjective.

  • Write-In. I finally gave up on this for a couple of reasons. One is that the logistics of another vote were tough, another is that I think by making other changes (weighting the nomcom to be community heavy and changing how we picked it for example) it was less important. I’m ok with not having this for this year, it’s something we’ll monitor to see if more needs to be done.
  • Values. I think we made some progress on this. People who lead PASS have to be vested in PASS, not just in the SQL community. That’s not unreasonable. Yet it’s hard to qualify and quantify, so we’ll see how it goes.
  • Less subjective. Sort of. If you’ve ever interviewed candidates for a position you know that hiring is subjective. Imagine being locked into a score sheet that forced you to hire the one with the top score even if you just knew they would be a bad fit (or someone further down would be a better fit). Unless you get the scoring system perfect it’s a recipe for bad hires. What we did was put a bit less weight on the oral interview, and change how the final list of candidates was picked. Without having tested it I like the approach and feel like it satisfies the heart of what I wanted to accomplish.

In the past the nomcom rated candidates in about 10 areas (education, leadership) on a scale of I think 1-5, and then used that to do the interview and again rated them in various areas. Then we totaled the scores and drew an arbitrary line about who was qualified to be on the slate and who wasn’t. Very very subjective, because we had no guidance on what was a 1 and what was a 5 for education (and it’s a different conversation on whether a Phd would do better than a high school grad).

What we changed to was a ranking system. No longer can the nomcom members just write in an arbitrary score, now they have to look at the whole  pack of applications, score them, and use that to help them decide on their final ranking. It’s not just picking who should be interviewed, it’s identifying right then who appears to be the #1, #2, and so on based just on their application. That’s important because we use it to put our values into the process, but we also leave some room for subjectivity. It’s also important because we’re using those rankings to decide who gets interviewed. Logistics are always a concern, and we know it’s not possible to interview everyone that applies – imagine we get 50 applications for example. We basically repeat the process for the oral interviews.

I won’t go through the entire process here, and I’ll grant you that when you look at it it seems complex, but I think it’s actually pretty easy to apply, and the upside of the revised approach is that we focus on comparing candidates to find the best candidates. It’s a process I’ve used before; score resumes and pick the top x, score interviews and pick the top x, and then a very subjective second interview of the top couple candidates to make the decision.

There is one other change that I think is important, the composition and selection of the nomcom. In all years past the Immediate Past President selected the entire committee. I don’t think that was a horrible plan, but it did mean that the nomcom had a tendency to reflect the views of the Board and the IPP. This year we’ll have nomcom members that are picked by PASS members, giving the members a deeper (and a majority) voice on the committee during the most crucial phase of the election.

Doing all of this gave me a new appreciation for how hard governing is. The tendency is to make a rule to fix every problem, a bullet point for every exception, and to root out subjectivity as an evil thing. It’s tough to know what things need to be rules and which things should be left to the committee. Too many rules and it’s just an algorithm, too few rules and its grounds for an argument or worse every time. Not easy stuff.

Will it all work? We’re hopeful. We tried hard, we tried for a sane and realistic strategy. We spent a lot of time looking at edge cases, trying to prevent the process from being abused or running amuck. I believe we should reconvene the original ERC post-election to see what worked and what didn’t, and based on that, we may need to form a brand new ERC to dig into specific areas.

My thanks to the volunteers who gave up a lot of time to help build the recommendations; Lori Edwards, Wendy Pastrick, Brian Kelley, Allen White, Bill Graziano, and Joe Webb.

Growing Leaders

Here’s a two part story(Part 1, Part 2 from Information Week by Larry Tieman you should read. There’s a great quote that to find the influential leaders within an organization “Look for the person with a line at their door”, something that rings true to what I’ve seen. There’s a related article about the five ,,core competencies of IT leaders that discusses growing from director to VP that’s hugely interesting as well.

Indexes Aren’t Always the Answer

There’s a tendency to think all problems can be solved with an index or two (or partitions, or other performance trick of your choice), but sometimes the answer is to not do the work to start with. Sounds Zen, but it’s often a matter of caching in the write place. It’s a technique I’ve used a few times in interesting ways, but maybe never quite to the scale mentioned in this article about about Accuweather. By doing something pretty interesting yet pretty simple out in front of the database they reduced their workload from 300 to 500…million requests per day.

It’s good to tune, good to have the right indexes, but tuning requires understanding the entire problem.

An Update on My Dell E6500

I bought my E6500 almost three years ago, giving away what was a perfectly good Latitude to trade up to the latest and greatest, including a 64G SSD. It’s been a good machine, runs Windows 7 without issue and even though I hibernate the machine daily, the SSD still goes on. I wasn’t at all sure that a 64G drive would be enough. I’ve had to prune files a couple times a year, but it’s been plenty for the stuff I use and need on a daily basis.

I thought about this only because my wife had asked if I wasn’t due to get a new machine, usually every couple years. I looked it over and it’s in pretty good shape, missing the “T” key where my baby girl decided to do some remodeling, and I just moved the tilde key over to replace the more often needed T. I looked a little closer and saw a small crack in the display bezel, not critical, and surely my fault from rough handling, but really it’s held u well.

I looked up the order to see when I bought it and saw that it was still under warranty and accident damage protection,just barely. I called Dell and a technician is supposed to come by this week to replace the screen and keyboard (they said I could keep the old keyboard for spares!). Neither is critical and if I had to pay for the repairs I wouldn’t,but with them I think I can go a while on this machine.

I spent about $2k on this one, and even then debated the merit of $2k versus going with a yearly spend of $600 or so on a lower grade use and discard plan once a year, but it’s worked out ok. I did a quick price check to see what I could get today, and the same machine (with 128G SSD, no 64G option, and 8G ram instead of 4g) was $1400 with the i5, a couple hundred more for the i7.

I’m in no hurry to buy, if anything I’d like to wait on one that can go to 16G or more to do better running VM’s, but at about $500/year for three years I like even better my strategy to buy a good machine and keep it versus an ok machine that may break. The math is fuzzy of course, I could well buy a $600 machine that would last 3 years, and even if it only lasts two I’m ahead. Still, in this case I’ll stick to what works, which right now anyway is my not quite three year old laptop!

Used Book Stores and Vacation Reading

I took a vacation recently (well, a really long weekend) and as I like to do, visited a few book stores during that time. One was Haslams in St Petersburg, in business more than 60 years and 30,000 square feet of books. I browed for a while and picked up one old book on wood, maybe 1910 or so, and I thought about how heavy it was, literally. Old books use that thinner paper and it’s denser, feels…serious. They also had a few display cases with ‘valuable’ books; first editions, signed copies, things like that. Being a book kinda guy I was tempted to buy one, just because!

The other was Paperback Palace,a more common store front type place at about 5000 square feet. Nothing really noteworthy about this one,a good no nonsense book store with mostly paperbacks, though it did one one set of shelves on the small 2nd floor that looked in danger of tumbling like dominoes!

I read a lot, sometimes I want to try something new, sometimes I want the equivalent of comfort food. For me that often happens when I see a book I’ve read before and think, hey, I want to read that again, or maybe I just see an author’s name and want to read something they wrote.

I was in a science fiction mood so I grabbed a few by Gordon R. Dickson, of which the best was The Pritcher Mass, which is hard to describe in a sentence. Let’s say it combines a doomed planet with a smart college student and a witch!

I browsed the Western section and saw Louis L’Amour, so I had to have one of those. I had read a good portion of the ones on hand, and finally picked Flint. Orphaned child gets raised by gunfighter, gunfighter gets kills, child goes to New York, gets rich, marries badly, and is diagnosed as being terminally ill. Decides to walk away from it all and go back ‘home’, winds up in the middle of a conflict, meets a new woman, and finds out he isn’t dying after all. This book in particular has a few ancillary characters, one is Milt, who works for the woman rancher owner that is in trouble. He’s the quiet, been around the block type, and toward the end is in the town saloon with most of the crew when the enemy starts to accumulate, a fight is brewing. One of the others suggest leaving before trouble starts, but Milt says “no, I ain’t gonna miss this”. A man after my own heart!

The other I grabbed was Golden Rendezvous by Alistair Maclean. It’s set on a cargo ship/super luxury liner and the main guy is the first officer. The voyage has gone badly, gets delayed, and once finally at sea odd things happen; a man disappears, another is killed, said first officer gets cracked in the head, and finally the ship is taken over and set on course to meet another ship. But they’ve picked the wrong ship, wrong first officer. He’s thinking and figures some of it out, has an interesting fight to the death in a cargo hold with loose crates sliding back and forth during a hurricane, and then carries the dead loser up a ladder to toss him overboard – and does it with a badly damaged leg to boot. Things happen, the bad guys seem to win, and as promised transfer everyone to the other ship, and it’s over. Or is it? The fun part of this book is what is unsaid. He’s not an ex-assassin, ex-whatever super dude, he’s the first officer on a cargo ship. Did I mention a missing atomic weapon too?

As I browsed and relaxed I thought about the simple joy of finding a book I last read, oh, 10 years ago, and that was written in 1962. As the world moves to ebooks will I be able to find those old but not quite famous books? Even if they all get converted and published, will I enjoy the discovery process as much? Will Amazon have an algorithm that shows me stuff I read long ago if I switch to vacation mode? Maybe. But somehow I doubt it will be as much fun as just wandering shelves of books.

PASS Update #57-Board Benefits, Travel, and Meetings

From time to time I get asked about how PASS Board members are compensated, so I thought it would make a good blog post, and I want to add to that some notes related to travel and meetings.

Board members are not paid a salary, stipend, or honorarium. We get free admittance to the PASS Summit, but not to the pre-conference seminars, and we do get a free copy of the Summit DVD’s. Our travel and host costs are paid to attend the Summit, but we pay for our own meals (note that most meals are included in the Summit admittance). PASS also pays our costs to attend Board meetings, typically 2-3 per year not counting the one we have at the Summit. For Board meetings we get reimbursed for airfare up to $500 (slightly more for anyone that lives outside the US), HQ books and pays for the the hotel room, and we get reimbursed for any rental car/tax needed (and we try to do ride share when we do). Per diem varies based on the city, but averages $50-$55 per day.

Non-Summit Board meetings are usually held at a hotel that can provide meeting space for 16 or so. Most of us try to arrive the night before the meeting so that we can do an informal dinner and get a head start on the work to be done. It’s also one of the key times when relationships are built. Breakfast and lunch are brought in so that we can meet start work over breakfast and continue right through the end of the day. Dinner on the first meeting night is usually required,and is optional for the second night (many want to get home that night). Dinner is usually one check and paid for by someone from PASS HQ or the senior Board member present. In practice that means I pay for the airfare,rental/taxi, parking, and maybe an airport meal or two up front, and then get the money back in 30-45 days.

For the past three meetings we’ve tried to co-locate with an event or chapter meeting (SQLSaturday Nashville, Dallas Chapter meeting, SQLRally Orlando), and going forward we expect to continue that. It’s incredibly important that we connect with our members, especially those that haven’t been to a Summit yet, and to do so as a group so that we can share the lessons and questions. We’re also trying to be a little more aware of the time/travel required for Board members. Just as I’m in favor of moving the Summit around, I’m in favor of moving the Board meetings around so that no one person suffers travel too much.

Aside from arguing for moving the meeting location around, my biggest complaints since I joined the Board in January 2008 have been the cost of the hotel, dinner, and the amount of work we get done (or not). We’ve had some dinners where the cost was higher than I though appropriate,, $40+ per person, and that’s not counting wine and drinks. We’ve also stayed at some hotels that I consider overly expensive, even with the discount that HQ negotiates.

Many will argue that as the Board is unpaid, buying them a nice dinner and a nice hotel room is the least we can do. Maybe there is some merit to that, but I think it sends the wrong signal to the members. Perception matters. I believe we should meet at a typical business class hotel (Embassy Suites for example) and live within our per diem. If want to hit Shula’s here in Orlando that’s fine, but we should do it with the $20 or left over after our share of breakfast/lunch costs, and then pay the difference ourselves. Ideally we’d all pay our own check and get reimbursed, but as a practical matter restaurants want a single check when you show up with 10 or more.

I will add that the hotels are a package; rooms, meeting space, breakfast, lunch, so it’s hard to tell what costs what. I know here in Orlando I can get a room at the Embassy for $130 or so before tax, breakfast is free (and cooked to order), and meeting space is $300 day (and often less if you buy lunch from the hotel).

I think anyone that volunteers to be on the Board deserves to stay at a clean and safe hotel, no different than I’d ask anyone that worked with me to stay at, and to get reimbursed for reasonable expenses. Outback is reasonable, Flemings is not.

I’ve been told that our average meeting costs $30-35k. That number seems high, but it may be realistic. We’re typically flying everyone in plus 3-5 people from HQ (which we need). That’s probably $1000 per person for airfare, hotel (2-3 nights), and related costs, and then we have the cost of meeting space and any incidentals. I’d like to see it cost less, and that may be possible, but I consider the meetings to be absolutely essential. I think we should publish the final cost so that the members can see it clearly.

The real work gets done at the in-person meetings. Our monthly phone meetings are one hour or less, for votes and quick discussions on ongoing business. Collaboration takes time, and works best when you get people in a room. Relationships get built over lunch and dinner, and you can’t do that over Livemeeting. Imagine managing a project for your business by getting the decisions makers together 3 times a year. It sorta works. Not smoothly, more of a lurching walk.

Meetings are hard at best. We tend to get an agenda published at the last minute, with my request being to see it at least 7 days prior so I can try to prepare. We used to do reporting by portfolio, which seemed like a waste of time (we can read it beforehand), but I think we’ve gone the other way to the point that we don’t hold each other accountable. Sometimes we don’t put enough thought into prepping a particular discussion (and that is not a PASS specific failing, it happens everywhere), sometimes we can’t decide how far into a discussion to go (are we in the weeds?), and sometimes we start an open discussion that turns into a great talk, or a free for all. Not all meetings can be about making decisions, sometimes they are about exploring ideas, and just as with sausage making, it’s something not fun to watch, but it’s how it works.

Sometimes we have good meetings, sometimes not. Meetings need a moderator and a leader, sometimes we have that, sometimes not. I think we try to be too structured in some places, not leaving enough time for groups to break off and talk. The most valuable time for me is dinner and after dinner, I still remember great talks with Greg Low right after I was elected that helped me understand how much I didn’t know about the international world of SQL and PASS.

What else? I typically get a free shirt or two each year. Conference calls are to a toll free number, and other calls I just do on my cell phone. I don’t get an allowance to buy anyone dinner or drinks or coffee, that’s all on me. I’d like to see Board members get an allowance for that, but maybe that starts to look like compensation and should be avoided.

I hope you as read this you’ll keep in mind the time investment we make. For me it’s conservatively 200 hours a year. 4.5 hours this week, more on some weeks, and hopefully less on others. I’m trying to do a solid two hours each week not counting in person meetings, and that is a struggle. I knew the deal going in and I’m not complaining about the time, or the lack of compensation. I’d like to see us spend a bit less for meeting, but I don’t think it’s going to be a huge amount, a few thousand a year. We try hard to be productive, we try to make good decisions, and overall we try to be good stewards of the resources we have available.

I’m planning to ask the Board to make some changes this year:

  • Require a Board vote to stay at any hotel that costs more than $200/night including all fees/taxes (excepting the Summit, where we leverage some free rooms), and ask HQ to target hotels that would be in the $150/night range
  • Schedule Board dinners at restaurants with an average entrée cost below $30
  • Disclose the final cost of each meeting outside of the Summit (where the costs are so blended I don’t think we could figure it out in a way that be worth doing)
  • Publish a publ
    ic version of our agenda at least 7 days prior to each meeting
  • Schedule all non-Summit meetings to coincide with a SQLSaturday, chapter meeting, SQLRally, or other similar event unless it’s an emergency
  • Schedule two full days for each non-Summit meeting, and schedule an after dinner workshop on evening of the first full day
  • Offer former Board members the opportunity to attend any PASS event at our true cost, as a way to keep them engaged and available

I’d like to hear your thoughts before I do so. I want Board members and prospective Board members to be treated fairly. I want our members to have a fair and positive perception of the money we invest supporting the Board. Are all of my ideas good? What would you change and why?