Hacker News Reader

Comment by fooqux | Hacker News Reader

Wouldn't that just favor apps that push trivial updates once a week that either do nothing, or just add bloat to software that already fulfilled its purpose quite well?

Personally, I like the idea that if I release an app that's actually feature complete and relatively free of bugs, it should be favorably ranked and not biased against due to it not being "new".

A great example of this is ConnectBot, which is a great ssh client for android. I can't remember if I've ever seen an update for it in the years I've been using it. Does this mean it should be ranked lower than JuiceSSH just because Juice is newer / has updated about once a month since it released? I'd say no.

Then again, I'm not a fan of the proposal of TFA for this same reason. Frankly, I think the store already gives us the tools to make our own decision by letting us filter by our device and by the current version.

View on HN

You don't have to do things stupidly.

You can say "Older than X, decay", "has feature Y, decay". This is the company that makes pagerank, I'm confident they could come up with a useful ranking and rating algorithm

True that Google came up with PageRank, and have some smart people thinking up ways to rank things I may want.

It's also true that as long as PageRank has been around, there have been people exploiting how it works for their own personal gain. See: My "update every week to boost my score" example above.

Personally, I sort of like the thought that if somebody releases a crap app and gets a bunch of 1-star ratings they are likely going to stick for awhile. Harsh, I know, but it does give companies some incentive to get it mostly right at first.

Speaking of, something I'd like to see is a developer ranking. Even something so simple as averaging all their apps' ratings would be helpful. Done right, it would help identify those soundalike apps, too.

>Personally, I sort of like the thought that if somebody releases a crap app and gets a bunch of 1-star ratings they are likely going to stick for awhile. Harsh, I know, but it does give companies some incentive to get it mostly right at first.

>Speaking of, something I'd like to see is a developer ranking.

It would discourage releasing early and often, something that's hard enough to do already. I think I'm finally starting to get it through my head/ego that I need to just say "good enough is good enough, release it into the world and get feedback instead of spending 3 more months 'perfecting' it." Knowing a single app with bad ratings in the beginning could bring me down for years would just be another excuse not to ship.

>PageRank has been around, there have been people exploiting how it works for their own personal gain

Yeah, they change the approach to attack exploiters. The idea there is One True Way that will never need defensive measures feels a bit like "We can engineer a solution!" type approach. It seems all that's really needed is ongoing effort to fight exploitation.

These are two different things:

1. Give more weigh to recent ratings in order to calculate the average rating. If there hasn't been any new rating, there is NO penalty for not having recent ratings. 2. Rank better the apps with more recent ratings. There is a penalty for not having recent ratings.

I agree with 1) (my proposal), but not necessarily with 2). Although it's true that ConnectBot has a lot of new ratings despite having been updated in October 2010.

Pushing out many updates in a short period of time with an exponential decay would have little effect. It's like a refilling a cup that's only down a few drops. Refilling a half empty cup (for example, an app that wasn't been updated in a year) would have a much bigger effect. That's how the math of an exponential decay works. The farther you get away from the start time, the more it goes down.

Reply on news.ycombinator.com