Server crash
CommonKitchen.com was offline for a few hours on Wednesday night and Thursday morning. Our server had hardware problems which took several hours to resolve.
As a small company, we’re obliged to rent server space, rather than owning our own hardware. This has advantages–for example, a hardware failure doesn’t require us to purchase expensive new equipment–but it also means that our server is in Dallas, while we’re outside Boston. We have to rely on our hosting company to fix certain problems for us, and while their support is generally excellent, miracles (like e.g. swapping entire hard disks into new servers) take a few hours.
We’re taking steps to minimize our downtime during events like this, but the technical details are really outside the scope of this blog, so we’ll explain them in a more technical post elsewhere.
Those cookies
We know you’re all drooling over the chocolate-chip cookies on the front page. If you click the photo, you’ll discover those are the Original Toll House Cookies from Marjorie Standish’s Keep Cooking–The Maine Way, as I made them last December. You can find them on Flickr; tag your cooking photos with “commonkitchen” to share them with the rest of us!
Why Common Kitchen on Facebook?
Yesterday I laid out the basics of our little starter Facebook application. What I didn’t do was explain why, given all the other things we have on our to-do list and the myriad details involved in staying in business, we took a day out to throw together a Facebook application.
As with anything, there are a lot of reasons. Certainly one of them is that Facebook is the darling of Silicon Valley this year; something happened this summer, and suddenly all kinds of people over the age of 27 have been signing up. (At 33, I think when I first signed up I was more than a standard deviation away from the mean age on the site. If that was ever true, it isn’t any more; in the past few weeks I’ve had several friends from my high school and college days add me to their networks.) If there’s one way to get by the “Cold Start Problem” I alluded to the other week, it’s getting your application in front of the millions of Facebook users.
It’s tempting to take Facebook at their word and use them as a baseline social network–to let Facebook handle all the “social networking” features, and just build on the networks and users they’ve already created. The problem there is that really using the information inherent in the network is only possible if Facebook is collecting the kind of network information we need. (This is not a problem for Common Kitchen, but it could be for others.) A related problem, going the other way: what kind of privacy implications would that have? When users add Facebook applications now, they pretty much give the application permission to see all the information they’ve added to Facebook. 90% of the time that’s harmless, but who wants to find the other 10%? Let’s dig up a different reason.
If we’re going to be useful to our readers, we need to keep them up to date about what’s happening in their network on our site. The feeds we launched this week are one step in that direction; the Facebook application, as it works so far, is another. Not everyone follows feeds, but if the information from a feed is also turning up on Facebook (which, as the comic above notes, plenty of people do follow regularly) the same purpose is served. And if that brings people back to our site more often, because they know there’s something for them here, then we’re both happy: we build a more useful application, and our users benefit more from it.
Finally, there’s the “toe in the water” reasoning. That goes like this: we certainly don’t give anything up by creating this application. And yet we may be able to take advantage of a really good opportunity later if we’re already there. So why not give it a try and see where it goes?
I don’t think any of these are single reasons why we built the Common Kitchen Facebook application, but it’s fair to say they’re all among the reasons it exists.
Getting on Facebook
On Tuesday, we jumped on a bandwagon and launched the Common Kitchen application on the Facebook Platform. Right now, it’s a pretty simple little thing: if you add the application to your profile, every time you review something on Common Kitchen, an item will appear in your Mini Feed saying so. If your friends have the application installed as well, they’ll see that story in their News Feed.
There are two pieces to this. Here, I’ll just explain the outline of what the application does and where we’re going with it. Later, I’ll address the question, why?
If you’re not already using Facebook, you’re not missing anything about Common Kitchen. We’re never going to provide functions on Facebook that aren’t present on the site itself. This application provides a way for people using Facebook to keep tabs on what their friends are doing on Common Kitchen; you can also do that by going to your profile page and clicking the “RSS” icon next to the “Activity in my network” headline.
To use the Facebook application, you need to be signed in on both Facebook and Common Kitchen, so if you’ve added the application you may find that every now and then Common Kitchen asks you to log in to Facebook. Then, when you post a review–of a new restaurant, a cookbook, a recipe you just tried–your friends on Facebook can see that in your feed. (If you want to avoid this for one review, you’ll see that there’s a checkbox with your review; un-check it, and we won’t tell Facebook about the review.) Second thoughts? Facebook provides a little “X” next to the news item so you can delete it from your feed.
Facebook also limits how many items from us will appear in your feeds; the limits are per-user over varying chunks of time, and there’s also a “fudge factor” they build in which essentially means we have little control over whether your review will be noted on Facebook at all. Most of the time it will.
So far, this doesn’t amount to a whole lot, but we’re still exploring the platform and figuring out where Common Kitchen and Facebook can overlap. We’d love to hear your ideas; we’re already looking at notifying you about which of your Common Kitchen friends are also on Facebook, and making it easy for you to add them there (if you haven’t already,) and vice versa. We’ll make it possible for you to share your favorite restaurants, cookbooks, and recipes in your profile. Any new functions will be added to the same Common Kitchen application, so if you add it now, you’ll see our upgrades as they happen.
If you have any good ideas about how to improve our application, or what other features we should add, feel free to get in touch! And if you’re wondering, yes, you can find Noah, Parker and Audrey on Facebook…
Food feeds
If you haven’t visited CommonKitchen.com in the last few days, you may not have noticed that the site broke out in feeds this week. Actually, even if you have, you may not have noticed; the only real change was that the site is now littered with little orange icons which say “RSS”. And, if you use a browser which supports feed auto-discovery, it will display its own icon on many pages that indicates that there are feeds to be had. (There’s probably one displaying right now; this weblog has had a feed all along.)
So what’s a feed, and why did we add them to the site?
To grossly oversimplify, a feed is a way for you to “subscribe” to a page on the site. This is not a subscription like an e-mail subscription; in fact, one of the reasons we added feeds was to avoid sending our users more email which they may or may not actually want. Instead, the feed’s address is kept by some software on your end: either web-based software like Bloglines or Google Reader, an integrated part of your web browser (called “live bookmarks” or something like that), or a separate desktop application like Vienna or FeedDemon. You can add feeds to pages like My Yahoo! as well. That program automatically checks up at that address now and then, and if we’ve added anything to the feed, it grabs the new item and lets you know.
As an example of a place we’re using feeds on the site, check out the restaurant browsing pages. As you go through, you’ll see a feed icon on every page. If you subscribe, for example, to the feed for Bath, Maine, you’ll be notified by your feed-reading software every time a new restaurant is added in Bath. If you look at the page for your favorite cookbook, you can subscribe to two different feeds: one which will alert you to any new reviews of that cookbook, and another which tells you about any new recipes from that cookbook which are added to our site. All of this without requiring you to visit the site every day.
Then, if you don’t want to get these updates anymore, you can just tell your software to stop checking that feed. You don’t need to tell us; we didn’t know you “subscribed” in the first place. You don’t have to give us permission to send you email, or make sure we get through your spam filter, or any of the other inconveniences which come with email nowadays. As Steve Jobs says, it Just Works.
We’ve gone a little beyond these basic feeds, as well. Yahoo! offers a service called Pipes which allows for combination and filtering of data sources like feeds; we created a Pipe which combines the four feeds which list the most recent cookbooks, recipes, restaurants, or food websites added to our site, and presents them as one big feed. You can think of that as a “What’s New on Common Kitchen” feed if you’d like. I’m also thinking about making a “Boston Area” pipe which combines the restaurant feeds for Boston, Cambridge, Somerville, Medford, Arlington, Newton, Brookline, Watertown, etc. into one big regional feed; you could probably build similar “regional” pipes for lots of other areas.
What could you do with our feeds? What other feeds should we add? Let us know, and we’ll get on it.
Update: We’ve added feeds to the Questions feature as well (remember that?) so you can use a feed to track new question additions, or track suggestions posted in response to one question.
Would you use Common Kitchen in your Kitchen?
Our friend Sam Guyer at Tufts sent us this link from the Sunday NY Times, about a new HP computer made for kitchen use. Both Noah and I were TAs for Sam, and he knows we like to hash through the implications of different applications.
The idea of a kitchen computer has been around for a while; I remember mid-’90s discussion of “internet appliances” which could be set up in the kitchen or elsewhere in the house; I think there may have been some faintly condescending implications about how computers were going to be so easy to use, even housewives could handle them. (This was clearly before someone noticed that the skill-set of a competent housewife overlaps closely with that of a CEO.)
HP seems to have really thought this one through, though. Rather than taking a computer, and finding some way to cram it into the kitchen, they’ve thought about what happens in the kitchen and built a computer to support that. My favorite feature: a touch-screen that can be cleaned off like a countertop. I could use one of those for the track-pad on my laptop, some days.
Being useful: Answering questions
Our discussions in the past week or so about what development needs doing on Common Kitchen has centered on the phrase “usefulness.” The idea comes from what Josh Porter of bokardo.com calls “the del.icio.us lesson”: that personal value to a single user comes before network value. (If you’ve never met del.icio.us, it’s a site which allows you to store your bookmarks on the web rather than on a single computer… and it also allows you to share any site you bookmark with your friends or the ‘net at large. Now owned by Yahoo!, Del.icio.us was an early pioneer in tagging.) Put another way, if nobody else used the site, would it still be useful?
If the answer is “no,” that doesn’t make the site useless, but it does make the site vulnerable to what Porter calls “the cold start problem,” the chicken-and-egg problem a site faces when it’s most useful when it has a lot of users, but needs to become useful in order to attract those users.
So the work we’re focusing on now is on making Common Kitchen useful even to a limited number of users, and one of the features we’ve added in this direction is Questions. Questions addresses the problem of missing information on the site by allowing you to ask for it. By way of example, consider this question Noah posted: he’s looking for a good pasta salad recipe, can anyone recommend one?
Or consider one of the elemental questions asked by everyone who’s ever planted a garden plot: what do I do with all this zucchini? (You have heard, of course, of the gardener who left a single zucchini sitting out on the back seat of his locked car? He returned to the car to find a window broken, and, rushing to see what had been stolen, found he now had two zucchini.)
Questions let you ask about recipes now, but as we improve the system, you’ll be able to ask for restaurant recommendations, cookbook (and cooking) help, or plain old advice. Take a look at the questions which are already here, and drop in a suggestion or ask a question of your own.
Dealing with Babel
As far as I can tell, only 35% of the online world speaks English as their first language. That percentage may vary a bit if we consider only the fraction of the world with the luxury of choosing their meals, but the non-Anglophone fraction is still too large to ignore.
Software support for multiple languages, however, is still a tricky problem. Thanks in large part to Audrey’s participation in our development so far, we’ve made an effort from the very beginning to build a site which can be easily translated and presented in multiple languages. We use the Globalize plugin for Rails, which means that every text string which displays on the site is fed through a function which checks for translations into the users’ preferred language.
We’ve made sacrifices for this. The biggest and most obvious one is performance; so far, Globalize doesn’t batch its translations for a given page, so each string generates another database request; any page will spawn several dozen independent translation requests, which isn’t terribly efficient. This is one of the biggest drags on our site’s performance at this point (though we are hunting down others and streamlining where we see opportunities to do so.)
Another, lesser sacrifice has been in plain linguistic friendliness. In an effort to keep our language spare and easy to translate, we haven’t indulged in the sort of colloquialisms (”Take me to the kittens!“) that you’ll find on other sites. Twitter used LOLcats as error messages for a while; we can’t do that if we want to be multilingual. So if our language is sometimes a little stilted… well, consider how translatable it is.
Obviously, we haven’t made the commitment to translating this blog, either. Maybe someday that will come as well.
On Julia, Julie, and “foodies”
JohnL reminded me that yesterday was Julia Child’s birthday–her 95th, if she were still alive. The timing was apt, as Alison and I had been listening to the audiobook of Julie Powell’s Julie & Julia, the book which emerged from The Julie/Julia Project, on our way to and from our weekend events.
The audiobook is with Alison, so I haven’t finished it yet, but I have gone back and done some reading in the archives of The Julie/Julia Project, and I have now been struck twice–once in the book, and once again on the blog, where they come up earlier–by Powell’s thoughts about Child’s place in modern American cooking.
“…I have had enough. Enough of the $40 olive oils and imported semolina flour and ‘please, Turkish oregano only.’ If I read one more dining guru gushing about ‘honest ingredients, treated with respect,’ I shall vomit, sir. And ‘Market Menus’? Don’t get me started. The well-meant ‘food revolution’ Alice Waters instigated some thirty years ago has metastasized horribly. The Victorians served Strawberries Romanoff in December; now we demonstrate our superiority by serving our organic, dewy heirloom strawberries only during the two-week period when they can be picked ripe off the vine at the boutique farm down the road from our Hamptons bungalow. People speak of gleaning the green markets for the freshest this, the thinnest that, the greenest or firmest or softest whatever, as if what they’re doing is a selfless act of consummate care and good taste, rather than the privileged activity of someone who doesn’t have to work for a living.”
Powell goes on to say,
“Julia Child isn’t about that.
“Julia Child wants you–that’s right, you, the one living in the tract house in sprawling suburbia with a dead-end secretarial job and nothing but a Stop-n-Shop for miles around–to master the art of french cooking. (No caps, please.) She wants you to know how to make good pastry, and also how to make those canned green beans taste alright. She wants you to remember that you are human, and as such are entitled to that most basic of human rights, the right to eat well and enjoy life.”
(You can read all of Powell’s pseudo-manifesto here.)
Now, I just came out of two years of graduate student living, and the dietary difference between graduate students and undergrads is that graduate students recognize that pizza is not the only food group–not that they can do much about it. I don’t have deep-seated notions about food, and I don’t have a manifesto to deliver about how I feel about food. (Generally, I’m in favor.) I get intimidated when people start throwing around the phrase, “foodie.” But if I was looking for a flag to rally around, I would probably head for Powell’s (and Child’s) a lot faster than many of the others being waved nowadays.
Are special-interest social-networking sites the next big thing?
It should be obvious by now that we hope the answer to that question is “Yes,” but until we find out, we’re happy to see other people, like PC Magazine’s Tim Bajarin, asking it as well. In his column, “The Future of Social Networking,” Bajarin points out that the concept of getting specialized information from interest groups is nothing new–it’s been happening as long as the internet has been around–but it hasn’t been brought up to the technology standard set by the “social networking sites” of the past few years.
Bajarin uses Scuba Diving magazine–colleagues of mine when Scuba Diving was owned by Rodale–as an example of a “vertical” site which would benefit from social networking technology, but it’s also an example of how “old media” companies struggle to keep up with the pace of change on the internet. When I was at Rodale, working on runnersworld.com and its associated sites, the Scuba crew were always the ones pushing the envelope, doing new and exciting stuff in their Savannah office while those of us back at the mothership in Emmaus struggled with the IT department to implement message boards that weren’t painful to use.
Explaining why you’d prefer a “vertical” site to a general one is pretty easy. If I want information on how to bake good brownies, I don’t want to have to weed out all the hits from vintage photography sites. If I want to read about marathons, I don’t want to hear about Marathon Oil, or the other places where the long footrace is (ab)used as a metaphor. Sticking to a single interest makes your searches and discussions more powerful because they are automatically more specific.