Inbound Thinking

This is the third post in a series about pain-reducing improvements I’ve made to my workstation. For more detail, read part 1 and part 2.

For the first couple of years working at HubSpot I kept my computer, monitor, keyboard, and mouse all on the same work surface. Our desks were very simple, no monitor mounts or keyboard drawers. At some point I realized that having the keyboard and mouse on the desk might be causing some of my pain. So I started looking for a good keyboard tray to add to my workstation. By moving the keyboard and mouse to a lower position I would have to reach less and my arms should hang naturally at my side.

I had a few requirements for the keyboard tray: I wanted a removable tray that would attach to my existing desk fairly simply; we rent the office furniture so I did not want to make any significant changes to it. I also was looking for a good range of movement: sliding, tilting, swiveling, etc. Additionally, I wanted the mouse and keyboard to fit on the same surface, I don’t like a wobbly separate mouse platform. After some research on Amazon and reading a bunch of reviews I settled on the 3M ATK65LE. It had all the range of motion I wanted, was big enough to hold keyboard and mouse on the same surface and claimed to be a tool-free installation.

As it turns out, the installation was not entirely tool-free installation. The mounting system works by hanging a metal bracket on the front lip of the desk and then attaching the underside with heavy-duty velcro. In practice the velcro sucks and the bottom frequently detaches from the desk. So I drove a couple of 1-inch screws into the underside of my desk and it worked great; the tray stayed in place and there was no visible sign that I modified the desk.

The adjust-ability of the 3M tray is great. It covers all sorts of combinations of tilt, slide and swivel. I would certainly recommend the tray to others. The only down side is that the surface is just a tad on the small side. The mouse and keyboard both fit but there’s limited room for mousing. Not a big deal if you have your mouse speed set high enough. The new working position did help with pain a bit. It kept me from resting my wrists on the edge of the desk and it allowed my arms to hang more naturally. Although it wasn’t a cure, it was a step in the right direction.

Sometime last year we started running into trouble with the HubSpot dashboard application for all of our customers. The dashboard was beginning to load slowly, very slowly. I investigated the slowness for a while, with help from my co-worker Stephen Huenneke, and eventually figured out that the problem was a collection of hundreds and sometimes thousands of items that were being loaded unnecessarily for the dashboard.

The dashboard’s purpose was to show customers what’s happening with their marketing: if they get new leads, new social media stories, new inbound links, etc. The first version of the dashboard had no paging and after 20 or so items it started to be a long page with much scrolling. Paging was added. But the first version of paging was all done in the freemarker templating language and, although we only displayed 10 items, all of them were still being fetched from the database.

It was obvious that loading all the items was a waste of resources, we only needed the first page of items. This was easy to do with MySQL, just throw offset and limit on the query, right? Well, not quite. Although it was easy to get the first page of items this way, it was not clear that there was a second page of items to be fetched. What if you have exactly 10 items? You get a full page of them and then it is not clear if we should show a link to the next page of items. We didn’t want to run a count query to find out how many total items there were since that is another database call and every call takes time, we wanted the dashboard to be a very fast load since it is the first page our customers see. We also did not care about how many total pages were available, we just wanted to show a single “next page” link.

After a few discussions, Stephen and I came up with what we called the plus-one paging model. To display a page size of 10 items, we fetched 10 + 1 (=11) items from the database, then, in freemarker, we looped over those items and displayed only the first 10. The 11th item was used to determine if we should show the link to the next page. If there were 10 or fewer items we knew there was no need for the next page. Getting a “previous page” button was easy: The “next page” link had a querystring parameter for page number. On the second and greater pages we showed a link to the previous page. The page number parameter was also used to calculate the offset for the other pages: page number * page size = offset (first page number = 0).

I have no idea if others already use the plus-one paging model, but I’ve found it to be pretty handy for situations where you need paging and do not need to know the total number of pages. If you find it helpful please let me know with a comment.

Ever since Facebook announced the ability to add a Like button to any webpage I’ve been mulling over the consequences. My thoughts solidified. And I have decided that Facebook is f-ing brilliant.

To understand why I believe that, start by looking at search. Search today (hereafter ‘Google’) is based on crawling the web and ranking pages based on a number of criteria: Page title, page content, inbound links, etc. These criteria may as well be arbitrary: There’s no guarantee that the top result is what I want, it’s just the result that plays the Google-game the best.

Facebook is going to change that game. When Facebook has a big enough list of what websites my friends like, they can shape search results accordingly. I’ll go to Facebook and search the web, not only Facebook. The results are simply ordered by number of likes. The results with likes from my friends will be first, but then they can show results of things that my friend’s friends liked, out to two or three connections away from me. The social graph will re-shape the way I think about relevance.

Often I’m in an unfamiliar part of Boston, or out of town on a day trip, and I want to find someplace to eat. I can pull up Google and see what’s nearby, I can pull up Yelp and see ratings, I can pull up Urbanspoon and have it choose for me, but none of these are what  I want. I want to know where someone I know has eaten and if they liked it. I really don’t care if a stranger rated this restaurant five stars, we probably have different tastes. But I know my friends and I know which among them I trust for a food recommendation.

Facebook can own this, too. Facebook has a unique opportunity to steal Foursquare’s thunder by making location a big part of the platform. Instead of opening Yelp or Urbanspoon I’ll just open Facebook and see what’s on the map that my friends like. It’s not limited to restaurants: If my friends are checking into a golf course or a nightclub I’ll see that too. It will change the way I make decisions.

This is Facebook’s game to lose. They’ve got hundreds of millions of users around the world. They’ve got a hundred million dollars of infrastructure. Facebook has already started taking the first steps with the Like button, and it’s f-ing brilliant. I’m excited to see what’s next.

Last night it took two and a half hours to watch LOST. The show is 45 minutes long. The extra hour and a half was spent coping with the utter disaster that ABC calls a video player. It got me thinking about internet video, and how backward the system is today.

For me, there are four ways to consume LOST: I can watch it when it airs on Tuesday night; I can watch it on ABC’s website the following day; I can download it from iTunes; or I can download it from a torrent. I don’t have a DVR so that’s not an option, although I wish it were. I could also probably watch the video on Ninja Video but I’d rather not deal with that site.

Watching the show when it airs is not always possible, this past week I was at a Red Sox game and in future weeks I’ve got other plans, so throw out the when-it-airs option. That leaves watching it on ABC.com or downloading it. I like being able to watch video without planning ahead, so the streaming option is preferable. But after the awful experience of using the ABC video player I’m not sure I ever want to use it again. There were so many bugs I was in shock. How does a big, rich company like ABC put out such a piece of crap? Probably as a way to drive more viewers to live TV where they have better monetization. I had to re-start the video no less than 10 times, each time trying to get back to where I left off and having to watch 30-second commercials just for the permission to skip ahead: ridiculous, I’m not doing that again.

So, by process of elimination, I’m left with downloading the show. I don’t mind paying to download and I’ve bought video on iTunes before. Unfortunately, it just doesn’t work. Apple definitely takes my money, but they botch the download. I’ve tried regular and hi-def video and in each case the download gets “corrupted” and needs to start again. I have to manually wipe out the partial download and start over; not fun when you’re downloading two gigabytes. After two days of attempted downloading I gave up and asked for my money back. Apple refunded me and also explained that this is a problem that happens all the time! I’m shocked that a company as successful as Apple and a store as profitable as iTunes can not reliably deliver the content they sell.

So what’s left? Downloading a torrent. Torrents are available, easy to find, fast and reliable. I can download video via torrent faster than I can from iTunes.

Why is internet video so backwards? I’m willing to pay, I’m willing to sit through commercials, hell, I’m willing to fill out a form with my contact info if you just let me watch the video reliably. The technology that is winning the video game is the underground, illegal source. The companies backed by real money and a professional development team can’t deliver as well as a small network of amateurs with video capture and ffmpeg.

It’s tough to feel bad for the video providers getting undercut by illegal downloads when their technology is so poor. They’ve got the power to win this war, but they have to get their distribution to work better than the illicit channel.

Sunday morning I drove up to Gilford, NH for Mother’s Day and managed to sneak in nine holes before visiting my parents. I played with Andrew, one of my oldest friends and an equal golfer. We went to Pheasant Ridge, a local course that we’ve played many times before.

Unfortunately, Sunday morning had some seriously windy weather — gusts of 30mph and  a wind chill of 25 degrees. Also, the course is at the top of a mountain so we were completely exposed to the wind, it definitely messed up a few shots for me. But, despite the not-perfect conditions, we had a good time. I played fairly well for my first outing of the season and Andrew and I shot the same score.

I’m really looking forward to playing more this summer and I’m already thinking about buying a new driver or some long-iron replacements.

Last month I wrote about ordering some new inline skates. They’re here and I’ve had a few opportunities to go out for a spin.

It has been years since the last time I went skating so I took it easy to start with and skated around a nice flat parking lot to get the feel back. Luckily it came back fairly quickly and I felt comfortable after about an hour. Krystal joined me for my second outing at the parking lot, she did pretty well for someone who has never skated before, but not quite ready for a lap of the Charles river.

Last week I had the opportunity to go for a skate before work, and did a nice circuit up and down the Charles. I was surprised how bad the sidewalks were for skating. I’ve run the route many times but while running I never noticed how uneven, bumpy and cracked the sidewalks are. Apparently the Boston side is better than the Cambridge side, so I’m going to try that next time.

I’m happy with the skates, they’re comfortable, smooth and just fast enough to have fun without killing myself. I’m looking forward to getting out more often this summer.

Ever since the weather improved in Boston I’ve been dying to spend more time outside. Krystal and I go for a lot of long walks throughout Cambridge and Boston. I also go running once or twice a week with the HubSpot Running Club. A few weeks ago I decided that inline skating (rollerblading, if you prefer) would be a great way to have fun outdoors and get some exercise.

Years ago I was heavily into aggressive inline skating. I spent a lot of time in skateparks and industrial parks, jumping, grinding, etc. I’ve calmed down a bit since. I still have my old Rollerblades but they are built for jumps and grinds, not speed. So I decided to buy some new fitness skates.

It was surprising how few places in downtown Boston carry inline skates. Maybe it is just too early in the year, but the only place that had skates for me to try on was City Sports. I went there this past weekend and checked out the new skates from Rollerblade and K2. The fit for both was basically the same, but the Rollerblades were a ten bucks cheaper. I don’t need the skates right away since I’m going on vacation next week, so I decided to not buy at the store and instead check online for a better deal.

I did a bunch of research between the Rollerblade and K2 product lines and decided that in most cases you pay a premium for the Rollerblade brand. The K2 skates were just as good based on the reviews I read. I wanted a skate just above the entry level K2 and picked out the K2 Moto 80. The skates have an aluminum frame, softboot with exo-skeleton and a nice closure system. I found a great deal from Skirack of Vermont and paid just under $120 for the skates with free shipping.

They will be arriving while I’m on vacation, but I’ll post an update here once I’ve had a chance to try them out.

Tupelo has been on my short list of restaurants to visit in Cambridge for a while now. So when a great deal popped up on Groupon I had to buy it. Krystal and I tried to get a table there a couple of times before but there was always an hour or longer wait. We finally got a table this weekend and it ended up being one of the best meals in recent memory.

Friday night, we managed to get out of the house early and arrived at Tupelo at about twenty past seven. After a few minutes hanging by the door we were greeted and told the wait was about 45-minutes. We grabbed drinks at the bar: I had an ABITA Amber beer which is brewed in Louisiana and was pretty good though nothing remarkable. After only a 20-minute wait we were seated at a nice window-side table near the bar.

The menu at Tupelo is dead-simple: Five appetizers, six entrees and a few each of sides, specials and desserts. For starters we decided to split the Watercress Salad with Roasted Pear, Goat Cheese and Candied Pecans. The food came out almost immediately, it was pleasantly prompt service all nice.  The salad was great, dressed with nothing but EVOO and lemon juice, a nice light start to the meal.

For entrees we ordered the Red Wine Braised Brisket and the PEI Mussels special. The beef was amazing. It was slow cooked, tender and flavorful. It was topped with a very light drizzle of freshly-grated horseradish sauce and was a perfect sharp counterpoint to the savory meat. There were also piles of a simple mashed potato and braised collard greens. The greens had a smoky barbecue quality to them that was fantastic. Overall it was a deliciously savory dish and I probably could have eaten two of them. I didn’t try the Mussels but Krystal was very happy with both the quality and quantity of them.

Tupelo keeps their southern theme for the dessert menu as well, offering up Brown Butter Pecan Pie and Red Velvet Cake. We opted for the Pecan Pie with a scoop of Bourbon ice cream, instead of the normal vanilla. The pie was great, not too gooey, packed with pecans, a reasonable portion. The Bourbon ice cream was great also and had enough bite to offset the sweet; they should keep it on the menu.

Overall, I was extremely happy with the whole meal. The restaurant is small, cozy and simple which makes the food stand out. Tupelo is the best restaurant I can think of in Inman Square and I’m glad to have it just a few minutes down the road: I will definitely be heading back.

I’m not sure why Bonfire is gone and Pairings has taken over, but the quality has dropped dramatically. Forgettable service, poor execution and scattershot menu combine to make Pairings a must-miss.

I booked a table for 7:15pm on Valentine’s Day; Krystal and I arrived to a nearly empty restaurant and a missing hostess — she also had to work coat check and was running back and forth.

The menu was all over the place: Italian firsts, Asian and Continental mains and French deserts, none executed well.

The Ravioli Carbonara was a single ravioli in a bowl of extremely salty,  thin, non-creamy sauce. The texture pushed the limit of al-dente, a bit too firm.

The Berkshire Pork looked nice but execution was off. The sausage accompaniement was overcooked to dryness. Broccoli sauce was underseasoned with awful flavor and texture. Several of the components did not work well together.

Creme Brulee was paper thin and so warm it became runny. The Lemon Tart was the best dish of the night: a nice balance of sweet and citrus.

Service was slow. Server was not knowledgeable about the menu or the wine list. It took nearly 20 minutes to get a check after final dishes were cleared, and that was only because we flagged down the manager.

The wine list was neither extensive nor impressive. Absolutely no vintage bottles listed, their best was a 5 year old bottle of Chianti. Pathetic for a restaurant called Pairings.

A big disappointment overall, especially at $60 per person. Based on the lack of turnout for Valentine’s I guess the word has gotten out. Hopefully when Pairings shuts their doors Todd English can bring Bonfire back.

This is the second post in a series about pain-reducing improvements I’ve made to my workstation. For more detail on the genesis of the problem, read part 1.

Some of the earliest pain I experienced while working was in my right hand. It didn’t take long for me to realize that  it was being overworked. Constant jumping between the mouse and keyboard and using the cramped laptop keyboard and track pad were causing a lot of trouble.

For years, I had used a trackball on my home computer; it was fast and pain-free. So I bought another Logitech TrackMan to replace my mouse at work. It worked great for a few weeks, but then I started to notice pain in my thumb. The constant thumb movement with this particular trackball became painful over the course of an eight or ten hour day. It also did not solve the problem of moving between the mouse and keyboard frequently. My right hand is dominant, so I was using it for most keyboard commands in addition to mousing.

I decided that since my right hand was being overworked, I would switch mousing duties to my left hand. I switched back to a normal mouse, since the trackball was right hand-only and  made a cold turkey switch to mousing left-handed. It was pretty awkward at first. I did not reverse the buttons, so I had to either hold the mouse diagonally, or use my middle finger for primary click and index finger for secondary. It made me hold my hand on top of the mouse instead of palming it, so I didn’t rest my arm weight on the mouse and that was good for my wrist.

Moving the mouse from right to left also gave my workstation a more convenient layout. Keyboard was center with attached number pad on the right and the mouse on the left. So the left-side mouse was closer to the main keyboard than my right-hand was. And there was less pain because I did not need to hold my arm out; I could just let it hang naturally at my side and would still able to reach the mouse.

It probably took two months to get used mousing left-handed. Most tasks were manageable early, but touchy things, like working in Photoshop, took a while to get right. Today I can be as pixel perfect with my left hand as I am with my right. Although switching to the left-handed mouse did not totally relieve all pain, it did help with my right hand and arm so I have not switched back.

In the next post of this series I plan on detailing some of the equipment changes I’ve made at my workstation.