Unscheduled downtime

Posted on Tuesday 2 February 2010

Our apologies. Mosuki failed to come back up after a reboot. We’ll have it back up as soon as possible.

Update: It’s back up! Sorry about that.

glyphobet @ 2:45 pm
Filed under: Mosuki
Why build a natural-language interface?

Posted on Sunday 29 March 2009

PotionFactory’s Better Software Through Less UI explains the motivation behind using a little bit of natural-language parsing instead of complex interface widgets. In their case, they built a parser for repetitions, but it was for the exact same reason — quick, easy-to-use UI — that we built a full-blown English-language date-time parser into Mosuki’s event-creation process.

glyphobet @ 10:18 pm
Filed under: Mosuki and User Interface
Artificial attention

Posted on Sunday 31 August 2008

FriendFeed’s new “fake following,” a feature that allows you to appear as someone’s friend without actually recieving updates from them, has gotten a bit of attention from Merlin Mann, Jason Kottke, who calls it “a little bit genius,” and Rex Sorgatz, who calls it “the most important feature in the history of social networks.”

What does this have to do with Mosuki? Well, although it’s astronomically unlikely that FriendFeed got the idea from Mosuki, we can’t help but point out with a grin that Mosuki has separated your public list of friends from the list of people who you actually share events with, and hear about events from, since our very first release, in 2004. Acknowledging the necessity of artificial attention is a critical insight, and it takes any social site one step closer to mimicking real social interaction and being a truly useful tool.

glyphobet @ 10:56 am
Filed under: Mosuki
Mosuki will now tell you when your favorite bands are in town

Posted on Thursday 20 March 2008

If you’ve wondered why the Mosuki blog has been quiet these last few months, it’s because we’ve been hard at work on what is probably our biggest new feature since we first launched. Mosuki will now tell you when your favorite bands are performing nearby, just like it tells you when your friends post a party, or when a concert gets posted to one of your scenes.

How does Mosuki know what your favorite bands are? Magic! Just go here. If you use iTunes, Mosuki will read your iTunes library and figure out your favorite bands. If you use last.fm or iLike, type in your last.fm or iLike username, and Mosuki will grab your favorite bands from them. Or if you’re old-school, and all your music is on vinyl and 8-track, you can just type in as many of your favorite bands as you want, on that page.

That’s all you need to do. Now Mosuki will include shows featuring your favorite bands in your daily emails, and you can add them to your calendar and share them with your friends just like any other Mosuki event.

As always, send us any feedback, comments, and bugs, and enjoy!

glyphobet @ 4:19 pm
Filed under: Mosuki and What's New on Mosuki
Relational databases and free-form data

Posted on Tuesday 18 March 2008

The Death of the Relational Database argues that relational databases aren’t good for applications that may add various additional relationships with new features, because the tables store not just the “objects” but the relationships between them.

In the early days at Mosuki, Jonathan implemented a table to store “certifications,” which are essentially directed, labeled edges connecting arbitrary rows, or objects. These edges can connect any two rows in any of several different database tables. And this power lets us implement major features that we didn’t plan for originally, like adding iCalendar feeds, discussion forums, scenes, places, and a myriad of object-specific user preferences to the site, with a minimum of schema restructuring. Often we can add whole features without touching the schema.

This table is also the largest, and most problematic speed-wise as it requires two joins instead of one (although Ross has got it pretty much under control now), which I think speaks to the original article’s point that relational databases aren’t the best choice for more free-form applications.

glyphobet @ 4:55 pm
Filed under: Mosuki
Google Reader’s two privacy mistakes

Posted on Wednesday 2 January 2008

Last month, Google Reader started sharing all “shared” items in your feeds with all of your Gmail contacts. If you’d shared, for example, a single item with just your spouse or lawyer or business partner, that item suddenly appeared for all your Gmail contacts. Since your Gmail contacts include everyone you’ve ever sent email to or recieved email from, this ended up sharing lots of items that people would rather not have shared with business colleagues, relatives, or casual acquaintances. One person even complained that Google Reader ruined their Christmas.

There are two privacy mistakes here. The first, as Scoble points out, is the absence of full-featured, granular privacy controls. It’s a mistake to assume, as Google did, that sharing an item in a collection and sharing the entire collection are the same proposition. It’s also a mistake to assume that sharing an item in a collection with a person is the same proposition as sharing the entire collection with them. Many web calendars made the same mistake early on, allowing their users to control the sharing of calendars but not of individual events.

The second privacy mistake is to ignore the discoverability of public data. Facebook made exactly this mistake in 2006, when they decided to aggregate all the changes that were made to your friends’ profiles, and show you those changes on your Facebook home screen. Everything they were aggregating was technically public information from your friends’ profiles, but people screamed bloody murder. The same thing happened with Google Reader; although the shared items were, technically, publicly accessible via obfuscated URLs, the only way to visit one was to receive the URL from a friend. The new sharing feature radically changed their discoverability.

When a piece of data is practically undiscoverable, users treat it as non-public. When that item’s discoverability is then increased, users react as if their privacy has been breached. (This is actually the flip side of security by obscurity, where a program assumes that undiscoverable means private. The reason software makes this mistake is because people do.)

Mosuki has had full-featured, granular privacy control from the beginning. And it lets you control both how easily your friends can discover your group-only events, and how easily anyone on the internet can discover your public events. And these are features that can, and should, be applied to any system that involves sharing potentially privileged data. Sounds like Google Reader users might like them.

glyphobet @ 3:08 pm
Filed under: Privacy
Free doesn’t have to mean unreliable

Posted on Tuesday 18 December 2007

Many new members of the “Web 2.0″ smörgåsbord seem to think that by placing the word “beta” in front of their product they somehow relieve themselves of the duty to provide a service which is as bug-free and available as possible.

The founders of Mosuki have always believed that providing a service for other people to use means taking on the responsibility to fix bugs and make sure the site is available — regardless of being “beta” or free of charge.

brainsik @ 7:28 pm
Filed under: Mosuki
Mosuki November update

Posted on Wednesday 28 November 2007

There’s a whole lot of little new features on Mosuki this month. We’ve redesigned the way that you RSVP for an event, and added a place to leave a quick, clever comment for your friends when RSVPing. The privacy settings for events, places, and other items are greatly simplified. We now have a published locations page, for concert halls, theaters, museums — any venue that regularly throws events. The event lists now have weekly editor’s picks and a list of most popular events. Plus there are a bunch of minor UI tweaks, improvements, and bug fixes.

Enjoy!

glyphobet @ 11:54 am
Filed under: Mosuki and What's New on Mosuki