Monday, September 19, 2005
Searchfox Beta Report: Day 1
So the first big change for day one's report is that I'm giving up on the whole scoring thing. You see, the whole login failure issue discussed in day zero came down to my not realizing that the Searchfox RSS reader login page is at http://rss.searchfox.com, and it's the Searchfox favorites login page is at http://www.searchfox.com. I've been thinking about trying to shunt some of the blame off to Searchfox, but that's a little tough to do; while one could make the argument that the layout at www.searchfox.com suggests that the username/password fields there apply to Searchfox RSS...well, it's a stretch.
In the end, that whole experience meant that I had to assign myself something like -5,000 points for being a dumbass, and +5,000 to Esteban Kozak of Searchfox for not sending me a final email along the lines of "thanks, dumbass -- that's two hours of my life that I'm never going to get back." At that point the scoring process just gets a little demoralizing for me, so out it goes.
On to the actual day one:
Getting Data In There
Exported OPML file from FeedDemon (the best RSS reader in existence). Started import into Searchfox. Waited...but not nearly as long as I'd expected, inasmuch as I'm subscribed to something like 350 feeds. Searchfox also gave me a nice message thereafter, listing the feeds that it couldn't locate or validate; since I didn't even recognize any of them, I shrugged and moved on. Very smooth.
It may just be due to the beta status, but it's worth noting that this is clearly aimed at people who have already drunk the kool aid: the subscribe/import section of the interface assumes that you've already got a stock of feeds you want to import, or that you'll add feeds manually -- none of this "here are some feeds to get you started" stuff here. There's a placeholder in the navigation for "suggested feeds," but not yet live.
Nice and clean, but a few personal gripes: first, while I'm not a big fan of frames, there are cases where they (or similar magic) just make sense. For me, a Web-based RSS aggregator is one of those cases. Since I have my 350 feeds roughly grouped into four folders, I only have about 100 feeds listed at any given time...but that's still a lot of scrolling up and down if I'm checking out items from feeds at the bottom of the list: scroll down, click on Scobelizer, pop back to the top of the page, read items, scroll down again to Techdirt, pop pack to top of the page, read items...you get the idea. While there's stuff I don't like about Bloglines, they're implemented so that my list of feeds and my list of items from the active feed scroll independently, which is a godsend.
Second gripe: feed items end up looking like headline-only summaries, and I hate headline only summaries. Clicking the little twistie next to the headline opens the item content, but I'd be much, much happier if I got something like 25 words of the post displayed by default. Headlines don't necessarily tell me whether I want to read an item. This seems like a potentially big issue for Searchfox, since it's deciding how to score feed items based on my interaction with those items. Just clicking the twistie and leaving the item text open seems to bump up its affility score pretty significantly...haven't yet tested how/whether I can train Searchfox to ignore or downgrade items that I read but don't particularly care about (short of using the delete function, which I'd like to avoid).
So far, so good: generally like Searchfox, and feeling like the personalization has really interesting potential. Some irritations, but still trying to figure out how much is just due to "that's not how I do things right now." More in a couple of days.