SSR: A very basic replacement for Google Reader

Google Reader is going away, which is too bad, because I love it. It is/was an awesome product with a lot of great features used by a lot of people for a lot of different reasons. I, however, started using it about a year ago, and for just one thing: to combine feeds from about 20 different websites into one long page. Stories come in and I read them and mark them as read, or save them for later. I occasionally used 'show all' (previously-read) and 'search' to find thing I've read in the past, but I never used tagging, sharing, or anything else. I loved that it was 100% browser-based for all the natural advantages that gives (like clicking on links and viewing content in the browser of your choice, working anywhere with no installation, etc.), although I was never happy with the way scrolling looked on an iPad -- it did this weird thing where as you scrolled, it would move every other line one pixel, then the other lines, resuling in a weird shimmering look. (And it showed up in regular mode in Safari and in "mobile" mode in Chrome. Very strange. And annoying, because on an iPad, I want to see the whole post, not just the title that I have to click to show the content.) Otherwise, I liked it and I was happy with its default behaviors, like "all collapsed on the phone; all expanded everywhere else."

Luckily, since my needs are few, I can more or less replicate what I like most about it pretty easily. I'm calling it SSR, for 'Super Simple Reader'. (And that's 'RSS' backwards. Ain't I clever?) If I were anything more than a second-rate PHP hack, I'd build it in Google App Engine (wouldn't that be funny) or Amazon Web Services, which you can use free for one year. It will be a good solution for me, and possibly for people like me: people who have their own webservers and enjoy installing LAMP apps on them. I'm starting with the MagpieRSS bundle and if I code long enough, I'll eventually recreate Tiny Tiny RSS. (Just kidding. TT RSS is awesome -- it basically makes your browser look like Netscape Mail stuffed full of feeds -- but a) I want mine to be very simple and b) I really want it to be optimized for mobile. TT RSS does have a mobile view but it's not what I want.)

Tiny Tiny RSS is awesome, but not what I want.

I don't have much time to hack -- usually just a couple hours here and there in a week -- but this isn't just a dream. I've actually started. I've got a few months, right? :-)

Random notes -- very much a work in progress...

• I'm building this with the most basic procedural code you'll ever see, but I'm making the database tables in Cake-like fashion: `id`, `created`, `modified`, and a general following of their field naming conventions. (Assuming they didn't change much in 2.x -- I last used Cake in 1.3.) So if you want to make this with Cake, you'll have a head start there.
• I don't know anything about the RSS spec. I'm just building this based on the output I get from the Magpie script. Fields are: title, link, description, pubdate, guid, summary, and date_timestamp. Since I'm not familiar with what format some of the fields should be in (if the spec says/requires that at all -- again, I am not starting with reading it) I'm just setting them to be VARCHAR(255). I am, at the moment, building my databases now. Once it's working, I'll publish the schema and I'll be open to suggestions for improvement.
• `guid` should (obviously) be unique but I'm going to handle that at the app level, and the database field will not be set to 'unique'. Feel free to change that in your own implementation. And if you want to scrap my `id` column and do everything with `guid`, go right ahead. I figure it's just better to have things like 'edit.php?id=5' in a URL instead of 'edit.php?id=http://some.site/some_article' -- or worse, 'edit.php?id=http://some.site/blog?story=blah'
• Are there standards for things like 'pubdate' and 'date_timestamp'? I'm just storing them as strings, but if they're enforced, I'd store them in a proper format. Oh wait... just remembered, RSS != Atom != whatever else there is. I'll just assume that Magpie does magic. Perfectly. :-)
• This is strictly quick-and-dirty. The app assumes everything is good -- for example, that the `url` field in the list of feed sources is populated. (Error checking? What's that?)

Questions? Comments? Any other feedback? Here I am.