June 29, 2005

Yummy and yummier

What do tags taste like? Dinnerbuzz may give us an answer. Think of the site as 43Meals: it’s a collaborative restaurant reviewing system where diners tag each restaurant review with tags like “outdoor”, “beer”, “excellent” etc. (Nothing has yet been rated delicious.) Each review also includes a rating.

The idea of using tags as a reviewing tool has lots of potential, since it offers the possibility of finding restaurants that meet a particular appetite for pizza, Persian or patios. Every city in the database gets its own page, which shows you which restaurants are being most frequently tagged in that city. And Dinnerbuzz gets full marks for offering XML feeds on each tag and city page: never again will you be the last to hear of that great new cafe in Menlo Park.

There are a few implementation hiccups: while you can search by tag (and even multiple tags) you can’t search by rating. But I’m not interested in finding a Mexican restaurant, I’m interested in finding a good Mexican restaurant, so I want to be able to limit the results of my tag search to restaurants with a certain minimum rating. And ratings alone aren’t going to pull me into a restaurant: I want to have some description that tells me why I might want to go, what I should order, and how I’ll feel the next morning. I’m guessing that the Dinnerbuzz gang was more excited about coding than eating because the reviews that are on there so far are mostly limited to tags and ratings.

The other challenge Dinnerbuzz faces is the challenge that folksonomies place on any site: until you get to a certain critical mass, the wide variation in how people use tags means that there’s an awful lot of noise and chaos. For example, I’m not sure how useful the tag vibe is in identifying possible dinner spots — yet it does produce a couple of results. But if I search on Indian — and who doesn’t want a curry at least once a week? — there’s only one result so far. This is a site that won’t be useful until there is a whack of people using it — and that could take a while. Meanwhile Dinnerbuzz may want to provide some gentle encouragement to tag restaurants with the information that folks are most likely to be seeking: ethnicity/genre; special features (patio, delivery, etc); and with a consistent set of qualitative tags like bad, mediocre, ok, good, excellent (right now this seems to happen strictly informally).

If Dinnerbuzz could hook up with structured blogging we might see bloggers helping to populate its reviews. I’d love to see Dinnerbuzz accumulate a useful set of reviews for Vancouver but I think that will happen fastest if it can integrate with existing review sites like VanEats.

If you like your yummies metaphorical rather than literal, check out Yummy, now in alpha testing. It’s a collaboratively tagged collection of PDFs on a range of topics, with a transparent business model: any of the PDFs can be printed and shipped to you by Print FU. If the PDF you want isn’t there you can upload your own.

Yummy might be an interesting way of finding PDFs, but heck, I can do that on Google. What would make me really happy is if Yummy had a set-up that let me store all the PDFs that are cluttering up my hard drive on their site, and made it easy to print from there without downloading the PDF in the process. That would provide folks with an incentive to upload and could quickly turn Yummy into an awesome virtual library.

Meanwhile, both Dinnerbuzz and Yummy remind me that how much I’d like folks to standardize the address syntax for tag-specific pages. To give you a sense of the variation, here’s how you’d find the tag page for “beer” on a few sites:
Dinnerbuzz: http://www.dinnerbuzz.com/tag.php?beer
Yummy: http://yummy.printfu.org/yum.cfm/search/beer
Flickr: http://www.flickr.com/photos/tags/beer

Could we all just give del.icio.us credit for its first mover status (and its very nice, simple URLS) and adopt the common syntax: http://www.yoururl.com/tag/whatever

I want to be able to go to any tagged site and know how to find the tag page I’m looking for. For that matter I want to be able to add any RSS feed for any tagged site to my aggregator without having to check the URL first, so again, let’s follow the del.icio.us example: http://www.yoururl.com/rss/tag/whatever

It may not seem like it matters SO much now but once we start building all of our daily information tools around tag subscriptions it will make life a heck of a lot easier.


  1. As far as apache or lighttpd-powered sites go, it’s a simple matter of rewrite rules to get the tag syntax like delicious, regardless of the original setup. It looks like the DinnerBuzz people just haven’t put any rewrite rules in…

    The real question is… how do we support more complex tag combinations? What if I want to search for results containing ‘curry’ and ‘indian’ but not ‘bangalore’?

    Should it be /tag/curry/indian/-bangalore ? /tag/curry+indian-bangalore ?

    Or how about searching for Thai OR Indian curry? /tag/curry/indian|thai ?

    Standardisation is nice, but it has to go further than just the basics, or the REAL difficulty with tag subscriptions will begin to rear its ugly head.

    Comment by unquiet — June 29, 2005 @ 1:44 pm

  2. The problem with complex tag combinations is determining grouping. For example, what if I want curry and either indian or thai? It might be curry+indian|thai, but do I mean (curry+indian)|thai or do I mean curry+(indian|thai). Does the parser need to know about grouping and if so have we exceeded the complexity levels that an average user is willing or able to support? This is my stopping point for adding complex functionality to Tagsurf.

    FWIW Tagsurf has been using the /tag/foo format from day 1 along with a lot of other nice friendly URLs.

    Comment by Anthony Eden — June 29, 2005 @ 2:15 pm

  3. Hi Alex,

    Thanks for your thoughtful and thorough comments. I really appreciate them.

    I just wanted to let you know that the “search by rating” feature has now been added to all global and local tag views. More is on the way!

    Thanks again,

    Comment by justinsmith — July 2, 2005 @ 2:28 am

  4. Profile: DinnerBuzz

    Company: Dinnerbuzz

    Launched: June 2005
    What is it?
    Dinnerbuzz is a new service that allows users to provide feedback on restaurants, bars and coffee shops. It leverages user-provided metadata (including tagging) to allow others to easily find an…

    Trackback by TechCrunch — July 3, 2005 @ 11:39 pm

  5. [...] I just came across Dinnerbuzz (catching up on my RSS after vacation; saw it via You’re It). Though the execution is iffy at best, the concept is close to what I’m [...]

    Pingback by BuzzMachine » Blog Archive » Made for the distributed world — July 17, 2005 @ 9:55 pm

  6. [...] king services it could become a very useful service over time. Links: Site About Tag Cloud You’re It (“It may not seem like it matters SO much now but once we start buildin [...]

    Pingback by TechCrunch » Profile: DinnerBuzz — January 21, 2006 @ 11:56 pm

  7. Yep

    There’s no place like ~

    Trackback by Emily — November 1, 2006 @ 8:13 am

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.