I’ll probably write up the exact details of implementation later on How I Use Things but I wanted to express my feelings about and behind this move and something so meta as that belongs here instead of there.
I’ve been using Google Reader for several years now, finding it an unparalleled application for managing feeds. However, I think it is seriously flawed in several ways and needs a lot of attention to make it useful in the long term.
The common response to a statement like that, in the open source world, is “OK then make it yourself.” I found that to be the appropriate reaction here and for the first Rails Rumble (original 2007 site is now completely gone, so you get a Wikipedia link instead) along with co-worker and friend, Chris Martin.
I named the new app Sluice and had some great features in mind. One of which was search for past feeds. Another was adding a social aspect to feeds and feed items that would allow you to mark something as publicly liked, among other things. There were some others, but those two things will have any Reader users right now saying “oh but they already do that”, which is true. Sharing just came out, almost two years later. And search was introduced two days before the competition.
On the one hand that’s a big blow, and indeed may have had a lot to do with why the project has been on indefinite hold since the contest ended (we had a very basic, but working feed reader with a pretty decent interface by the time the weekend was up). On the other hand it was very encouraging for a user of Reader. It told me that smart people were actively working on the application to improve it, and they were likely listening to other users who saw the same flaws and unmet needs that I saw.
Unfortunately that evolution has not continued. The sharing features that I had not planned to implement until well after more important needs were addressed are now in play with Reader, but do we have any way to filter feeds based on keywords? Prevent duplicate entries? See links to possibly related entries? It seems to me that they’ve only plucked the low-hanging fruit (they are a search company after all, how hard could adding that have been?) and are not truly putting the effort into evolving the core solution of Reader: consumption and management of feeds.
After once again surpassing the 1000 unread entries mark, I lost my patience with Reader and moved forward with the unholy union that I’ve been dreading since before conceiving of Sluice. I’ve wed Google Reader with Yahoo Pipes.
This is nothing other than an act of desperation. I’m a practitioner of Inbox Zero and having all of my mail processed, organized, and dealt with is an unbelievable relief. The same is true of feeds. The more unprocessed items I have in Reader, the more daunting, heavy, and unworkable it is. Time-sensitive items are lost in the flood of news about Ballmer’s latest clumsy attempts at edging around Linux in the marketplace, or everyone and their grandmother’s opinion of Snow Leopard. Valuable insights into topics relevant to my work and studies are plowed under by countless redundant 5/10/15/30 top/best/most awesome tips/tools/sites/snippets/tutorials for newbies/designers/developers/people with nothing better to do.
So I finally did it. Or I’m in the middle of doing it, anyway. As I process down all of the unread items in each feed, I’m moving it to Yahoo Pipes, where it is being filtered by keyword (“windows”,”erlang”,”nettuts”, etc.), united with the rest, passing through a deduplicator based on title and/or link, and finally being spit out in one big feed.
The downsides to this are so steep that it’s taken me over two years to execute on it:
-
Feed addition. When I come upon a new feed I want to track, I have to decide whether it’ll require filtering and deduplication. If so, it has to be manually copied into Pipes, rather than the two-click addition to Reader.
-
OPML Who? I can no longer generate a list of the feeds I track in Reader.
-
The tinyurl dilemma. Because of the obfuscation layer, I can’t participate in the social features added to Reader.
-
Gardening. I will essentially have to abandon the idea of dropping old feeds that are no longer positively contributing, unless I do outside research to determine what the feed is injecting into my Pipes rig.
-
Reader machinery. I’m uncertain how this will turn out, but I fear that because I will be the only one pulling this feed, the Reader-feeding machinery will not poll my Pipe feed often for updates.
Once upon a time it didn’t seem like a big deal to just click any article I want out of my way to mark it read, and move on to the stuff I wanted to read. But as I’ve added and removed feeds over time, I’ve found a core of them that are indispensable, however they take so much effort to maintain that no time/interest is left over for actually reading the items that do have value to me.
Optimally, the Reader team will surprise me with some features that handle the keyword and duplication filtering that will make this manageable for me, causing me to do the whole process over again, in reverse. I’ll be happy to do it when that day comes. Until then, hard compromise.