Querki's first Use Case is us sending out a save-the-date link for our wedding -- the British relatives need lots of notice, so we set a deadline of Jan 1. That's meant a lot of programming for me over the past six weeks, and a non-trivial amount of server configuration for Aaron on the IT side.
We're not done with the project yet -- we still have to write the text and I have loads of CSS hacking ahead to figure out how to make it pretty -- but I think, knock on wood, the technology is ready. Here's the prototype page, and here's the same thing without the surrounding chrome.
I'm not kidding myself -- at this point, it's basically just a wiki with one or two unusual features (like CSS support, and chromeless support for paid users). But everything *thinks* in a "Querki-ish" way. The photo isn't just an image, it's a full-scale object in the database; so is the stylesheet. The stylesheet has its own page, with display text separate from the CSS so that there is a place for usage notes, and is an instance of the Stylesheet Model, which sets up the typical properties automatically. The page itself uses the Stylesheet Property to point to the Stylesheet Thing. It's a simple illustration of the key notion behind the system, that it's as consistent as possible: everything is a Thing, every Thing can has whatever Properties it likes, and it all hooks together into a nice little weblet.
And all of this took about three minutes to set up, start to finish, which is kind of the point of the exercise. Ease of use is mediocre at this point, but much of the middle of the year will be focused on that.
So -- progress. I'm probably going to spend January adding more key features, and then much of February stepping back and fleshing out all the non-feature stuff that needs to be done. (Everything from getting clustering working to building the automated test harness to incorporating.) With any luck, we'll be ready for folks to start playing with it in April...
We're not done with the project yet -- we still have to write the text and I have loads of CSS hacking ahead to figure out how to make it pretty -- but I think, knock on wood, the technology is ready. Here's the prototype page, and here's the same thing without the surrounding chrome.
I'm not kidding myself -- at this point, it's basically just a wiki with one or two unusual features (like CSS support, and chromeless support for paid users). But everything *thinks* in a "Querki-ish" way. The photo isn't just an image, it's a full-scale object in the database; so is the stylesheet. The stylesheet has its own page, with display text separate from the CSS so that there is a place for usage notes, and is an instance of the Stylesheet Model, which sets up the typical properties automatically. The page itself uses the Stylesheet Property to point to the Stylesheet Thing. It's a simple illustration of the key notion behind the system, that it's as consistent as possible: everything is a Thing, every Thing can has whatever Properties it likes, and it all hooks together into a nice little weblet.
And all of this took about three minutes to set up, start to finish, which is kind of the point of the exercise. Ease of use is mediocre at this point, but much of the middle of the year will be focused on that.
So -- progress. I'm probably going to spend January adding more key features, and then much of February stepping back and fleshing out all the non-feature stuff that needs to be done. (Everything from getting clustering working to building the automated test harness to incorporating.) With any luck, we'll be ready for folks to start playing with it in April...