[HN Gopher] Two months of feed reader behavior analysis
___________________________________________________________________
Two months of feed reader behavior analysis
Author : AndrewStephens
Score : 27 points
Date : 2024-08-03 11:34 UTC (11 hours ago)
(HTM) web link (rachelbythebay.com)
(TXT) w3m dump (rachelbythebay.com)
| kibwen wrote:
| _> rawdog /2.24rc1. Behavior is spot on. More like this, please._
|
| I'd like to see a description of what the proper behavior is in
| this context. The OP uses terms like timing, pacing,
| conditionals, and unconditionals in a way that makes me think
| that these must be well-defined jargon in the context of RSS, but
| I don't see these in the RSS spec.
| gnabgib wrote:
| See Rachel's past posts, there's been several on the topic
| (although the service/scoring doesn't seem to be linked
| anywhere).
|
| [The feed reader score service is now
| online](http://rachelbythebay.com/w/2024/05/30/fs/)
| aquova wrote:
| Can someone explain how to parse this article? There's a number
| of similar entries for each client. I can't tell how to interpret
| these complaints.
| butz wrote:
| Sad to see that so many feed readers are unable to solve polling
| problems. It should be in their interest to make least requests
| possible. Especially the ones hosting multiple tenants that are
| very likely subscribing to same feeds.
| JustARandomGuy wrote:
| _Unread RSS Reader. Godawful poll timing. 6103 requests in 52
| days is about one poll every 736 seconds _on average_, but they
| 're hugely spread out. WTF? Put it this way: the list of unique
| intervals (nn seconds, nn minutes, ...) is four pages tall on my
| web browser._
|
| Not entirely sure what the criticism here is other than polling
| on average every 12 minutes seems a little excessive at best. Why
| does in matter it the intervals are a bit wonky? I could think of
| many reasons why: maybe the poll intervals are smaller during the
| daytime and more spread out over the night to optimize for
| reading conditions, etc
| Twirrim wrote:
| Disappointing to see feedly in the misbehaving bucket. It's an
| online service, you'd have expected to get this right
| pipingdog wrote:
| It's important to consider that an online service's poor
| behavior is amortized over it's user base (or all of the users
| who subscribe via the service's local copy of a given feed), so
| it _could_ be worse, if all those users were fetching the feed
| themselves.
|
| It's also the case that a service wants to ensure they have the
| freshest copy or impatient users could bail somewhere else, or
| just do it themselves.
|
| But, there's certainly an opportunity for a service to perform
| analysis on feeds to see the rate at which they're likely to
| have more content, as well as take cues from metadata.
|
| At the end of the day, RSS isn't a protocol, and feed providers
| are just as wild west as consumers.
| jerlam wrote:
| I'm disappointed in Feedly here, as I've been on their platform
| since Google Reader was shut down - over ten years. And their
| script is still at 1.0.
|
| Rachel's complaint about Feedly's overzealous polling also
| contrasts the experience of author John Scalzi, where his blog
| was simply getting ignored by Feedly:
|
| https://whatever.scalzi.com/2024/08/02/the-feedly-issue-appa...
| kkfx wrote:
| Personally I choose tt-RSS for a reason: human quickness going
| through posts. I can have all my feeds arranged, just a line per
| entry, scrolling mark them as read, opening one open a side panel
| without touching visually the current post and threads list,
| being on a desktop with a 21:9 screen it's a perfect mach. I've
| used Miniflux before but it's less quick skimming posts, opening
| one, come back to the compact list etc.
|
| Before, but very before in the past, I've tried few desktops one,
| and the final choice was a Java/SWIG one RSSOwl, heavy but
| effective enough compared to others. I've also used for a period
| elfeed (Emacs) but it's too slow to read feeds for me, Emacs
| offer UI for focused read, while for a large amount of low
| importance posts I need something allowing a very quick pass even
| if something get lost.
|
| A thing I miss in ALL feed readers I've tried so far is real-
| world filtering abilities like fuzzy-matching titles to
| reasonably show only one per kind of news (let's say you follow
| many newspapers and nearly all report the same news about an
| earthquake somewhere, there is no point in seeing let's say 12
| posts on the same event), eventually offering a button to show
| all the matched if I want to go through them. Another is
| historical analysis, let's say every years there are wildfires
| around the world, I'd like to see from news I've read if they are
| almost the same of the last year or more, if they start to appear
| earlier and last longer etc, it's still fuzzy keywords matching,
| nothing so hard, but still absent in all, I imaging I'm one of
| the very few interested in such automation to use feeds as a
| personal aggregator. Gnus with scoring offer something similar,
| but well, it's too slow to really skim things, and easy to break
| as well.
| impure wrote:
| Looks like I need to update my reader to use the If-Modified-
| Since and If-None-Match headers
| trekz wrote:
| Not surprised to see Reeder in there. It's a great app for Apple
| users. But that app can bring a website to its knees with how
| aggressive it is.
|
| I can see in my logs that it constantly makes over ~20 requests
| to different RSS feeds on my domain, all in the exact same
| millisecond. Happens multiple times a day. And it appears to
| rotate IPs. Scary... Tried reaching out to the developer about it
| twice, but they never responded.
___________________________________________________________________
(page generated 2024-08-03 23:01 UTC)