Existing users, 
 FlyFi iTunes Helper - 1.0Community and recommendation system for iTunes. |
|
||||||||||||||
|
|||||||||||||||
Feedback Summary:
| This Version: | |||||
| Overall Rating: | Not rated (0.0) | Features: | Not rated (0.0) | Support: | Not rated (0.0) |
| Ease of Use: | Not rated (0.0) | Quality / Stability: | Not rated (0.0) | Price: | Not rated (0.0) |
Key to Types of Feedback:
Reviews 
Troubleshooting 
Usage Tips 
Developer Notes 
Commentary 
Featured Reviews
eh, iLike is far superior 
- Version: 0.98.3, 1/10/2007 06:09AM PST
(0 of 1 users found this comment useful)
rastaclem
...and just a better product. The website for iLike ties in very slickly with the sidebar client, something that Goombah isnt able to do. I think the idea is neat, but could be more polished and better integrated into iTunes.
Slow, buggy but interesting - Version: 0.9606, 8/24/2006 05:31PM PST
(1 of 1 users found this comment useful)
machspeed
rather slow, but great free music selection, and i actually really loved the #1 recommendation it gave me. senegal fast food. gotta work on making it more mac friendly and speedier.
Goombah is pretty cool - Version: 0.9606, 7/17/2006 05:37AM PST
(1 of 1 users found this comment useful)
MPolacek
I found this a few weeks ago and tried it. It's pretty cool. I really like a lot of the free music that it recommends to me. There's new music every week. It's also fun to browse other users collections.
Update from Gary Robinson's Blog - Version: 0.9606, 7/10/2006 01:02PM PST
MMooney
Well, the day is finally here. We have been revamping Goombah in a big way for a long time. Over the last week or so we've been rolling out the first public updates to ourGoombah beta since version .613 more than a year ago. (It's for Mac and Windows.)
From the outside, I'm sure it may have seemed that we'd slowed down or even moved onto other things. Wrong! The company grew, and we've been pushing hard on all development fronts. The user interface has been revamped, cool features have been added, and the recommendation engine has had major improvements. We just wanted to wait until things were really right before we put it out again.
One of the great new things: we are making available free music that is downloadable to any device including iPods. Every week we have Free Music Friday, when we'll make new music available from great acts such as Richard Thompson, 50 Foot Wave, Aimee Mann, and Janis Ian, all of whom have tracks available now on Goombah.
With regard to the recommendation engine improvements, sometimes I'm actually kind of stunned at how great it's working at this early stage when we're really only coming out of the door. It's based on collaborative filtering, so as our user population grows, it will only get better and better. (That's a reason to tell your friends about us, hint, hint!)
Goombah is still in beta testing, so we don't guarantee you're not going to run into bugs. We DO guarantee that we'll greatly appreciate any bug reports and put a high priority on fixing them. Overall, however, it's stable and works. There's no reason not to use it now.
From time to time I'll be offering tips in this blog.
Tip #1: Tune List Settings for My Library.
Default Goombah recommendations are based on all the music in your collection. But for many of us -- myself included -- a lot of the music in our collections is there for sentimental reasons or because of just checking stuff out. It doesn't necessarily all represent our current tastes. So, we'd like to tell Goombah not to use it for recommendations.
To do this, click My Library in the table on the lower left, and click the tab for List Settings. You'll see a check box next to each track. Uncheck all tracks from the artists you feel aren't representative of your current tastes.
To do this quickly, you can select a number of tracks at once by shift-clicking. Then right-click (or control-click if you have a Mac with a single-button mouse), and select the option to Uncheck.
Tip #2: Make new searches.
My favorite thing to do is click the "+ New List" button in the lower left, and enter a list of artists. Goombah then finds other members who are unusually similar to you in liking those same artists, and generates recommendations for other artists who have something in common with all the ones you entered.
It works really well. Just try it.
From the outside, I'm sure it may have seemed that we'd slowed down or even moved onto other things. Wrong! The company grew, and we've been pushing hard on all development fronts. The user interface has been revamped, cool features have been added, and the recommendation engine has had major improvements. We just wanted to wait until things were really right before we put it out again.
One of the great new things: we are making available free music that is downloadable to any device including iPods. Every week we have Free Music Friday, when we'll make new music available from great acts such as Richard Thompson, 50 Foot Wave, Aimee Mann, and Janis Ian, all of whom have tracks available now on Goombah.
With regard to the recommendation engine improvements, sometimes I'm actually kind of stunned at how great it's working at this early stage when we're really only coming out of the door. It's based on collaborative filtering, so as our user population grows, it will only get better and better. (That's a reason to tell your friends about us, hint, hint!)
Goombah is still in beta testing, so we don't guarantee you're not going to run into bugs. We DO guarantee that we'll greatly appreciate any bug reports and put a high priority on fixing them. Overall, however, it's stable and works. There's no reason not to use it now.
From time to time I'll be offering tips in this blog.
Tip #1: Tune List Settings for My Library.
Default Goombah recommendations are based on all the music in your collection. But for many of us -- myself included -- a lot of the music in our collections is there for sentimental reasons or because of just checking stuff out. It doesn't necessarily all represent our current tastes. So, we'd like to tell Goombah not to use it for recommendations.
To do this, click My Library in the table on the lower left, and click the tab for List Settings. You'll see a check box next to each track. Uncheck all tracks from the artists you feel aren't representative of your current tastes.
To do this quickly, you can select a number of tracks at once by shift-clicking. Then right-click (or control-click if you have a Mac with a single-button mouse), and select the option to Uncheck.
Tip #2: Make new searches.
My favorite thing to do is click the "+ New List" button in the lower left, and enter a list of artists. Goombah then finds other members who are unusually similar to you in liking those same artists, and generates recommendations for other artists who have something in common with all the ones you entered.
It works really well. Just try it.
Recommendation Quality - Version: 0.521, 3/5/2005 08:25AM PST
(1 of 1 users found this comment useful)
garyrog
Three things everyone should keep in mind for now:
a) The recommendation quality is dependent on the number of users, because it works by finding similar users to you. It's essentially supercharged word-of-mouth, supercharged because we use mathematical processing and a lot of computer horsepower to find the VERY best matches to you, so that you are extremely likely to like the same things your best matches like. BUT: if your tastes are non-mainstream, the probability that there will be matches to you is based on the number of people who have signed up already. And because we are still in fairly early beta test mode, we have done almost no publicity for it. The publicity has been a blog and a couple of mac version sites. We are using this period to get the code right, refine the UI, etc.
b) The bottom line is that it does not make sense to judge us on recommendation quality... YET. We're going to put our tails on the iine and say that Goombah will have the best recommendations in the business. Our software is built from the ground up to handle huge numbers of users with no deterioration due to scalability issues. That's what we're really about. It means that when we have lots of users, we WILL find your VERY best matches, according to mathematical processing based upon a key optimality theorem. At that time, the recommendations will be fantastic. But right now, while we prepare for the growth we are planning, we should be judged according to UI issues, stability, etc. We simply cannot produce those fantastic recommendations until we get more people, and we aren't trying to get a lot of people now because we're beta-testing and refining the UI.
c) The above being said, we will be posting an update before long that should make the recommendations significantly better than they are now. The algorithm in the software currently assumes we have a lot of users, because our whole is to harness that power of that kind of large community. But there are simple things we can do to minimize the negative impact of the fact that we don't have those people yet. The coming updates will start to include such processing so that recommendation performance is substantially better during this testing period.
We'll post developer notes here regarding changes that involve recommendation quality as they come.
a) The recommendation quality is dependent on the number of users, because it works by finding similar users to you. It's essentially supercharged word-of-mouth, supercharged because we use mathematical processing and a lot of computer horsepower to find the VERY best matches to you, so that you are extremely likely to like the same things your best matches like. BUT: if your tastes are non-mainstream, the probability that there will be matches to you is based on the number of people who have signed up already. And because we are still in fairly early beta test mode, we have done almost no publicity for it. The publicity has been a blog and a couple of mac version sites. We are using this period to get the code right, refine the UI, etc.
b) The bottom line is that it does not make sense to judge us on recommendation quality... YET. We're going to put our tails on the iine and say that Goombah will have the best recommendations in the business. Our software is built from the ground up to handle huge numbers of users with no deterioration due to scalability issues. That's what we're really about. It means that when we have lots of users, we WILL find your VERY best matches, according to mathematical processing based upon a key optimality theorem. At that time, the recommendations will be fantastic. But right now, while we prepare for the growth we are planning, we should be judged according to UI issues, stability, etc. We simply cannot produce those fantastic recommendations until we get more people, and we aren't trying to get a lot of people now because we're beta-testing and refining the UI.
c) The above being said, we will be posting an update before long that should make the recommendations significantly better than they are now. The algorithm in the software currently assumes we have a lot of users, because our whole is to harness that power of that kind of large community. But there are simple things we can do to minimize the negative impact of the fact that we don't have those people yet. The coming updates will start to include such processing so that recommendation performance is substantially better during this testing period.
We'll post developer notes here regarding changes that involve recommendation quality as they come.
Most Recent Replies: View All 3 Replies
- Recommendation Quality
Nothing fantastic - Version: 0.521, 3/4/2005 02:49PM PST
narasent
Not too impressed with this app. Almost none of the recommendations were useful - it either brought back recommendations of songs/albums already in the library, or many that I would not purchase or listen to. I don't recommend this app, it would be a lot easier and more beneficial to just go to iTunes or Amazon.com and just browse their recommended music. (especially since Goombah takes you to the iTunes music store anyway when you want to check out some of the music it suggests)
Most Recent Replies: View All 1 Replies
Too Big! - Version: 0.502, 10/14/2004 02:54AM PST
(1 of 1 users found this comment useful)
Digital Fruitcake
I'm out here in the sticks on a 56k link, so even though I'd love to try it out I can't justify the painfully long download time I'll get from the near 30 meg dl. For now I'll be sticking with iScrobbler.
Most Recent Replies: View All 1 Replies
Whoops - Version: 0.21, 12/11/2003 03:24PM PST
freaktheclown
Forgot to rate this. Definately 5 Stars...there's not that many people, but the app itself is great. Look forward to future updates!
Did You - Version: 0.21, 12/11/2003 03:15PM PST
freaktheclown
respond to the confirmation email? Maybe that's why it won't connect. Also, it does take like 20 seconds or so to connect, so just wait and it'll go.
Worked great for me...now we just need more people.
Good work though! Interface is nice and intuitive.
Server better now - Version: 0.21, 12/11/2003 10:19AM PST
garyrog
We just made a change on the server side that reduced database CPU usage by 80%, and we think this was the main bottleneck. So server connectivity should be better now.
We're monitoring this situation very actively. Remember that the iTunes Music Store itself practically shut down the first day or so due to the unexpected load, but then was fine. (And frankly we've been ramping up a good deal faster than we expected.)
So Big Bass Drummer and Tamagosan2 (and anyone else running into server problems) -- now would be a good time to try again.
