Testing user profiles for ReadBurner

I’m testing kind of user profiles for people who share things on ReadBurner.

Instead of immediately  linking to the user’s linkblog I link to a site where you can see what the user has shared. Since you can see everyone who shares item, you now can more easily find people who share same stuff. Here’s my profile for an example.

I think this feature leaves a lot to be desired, but first a major design update is getting more and more important if I want to keep up adding new features without completely overloading the site.

Another thing that I totally forgot is to implement a next button if you wanna see more items. Expect this also after a design revamp.

DreamHost sucks. Share that.

Not just because of that billing fauxpass, but they just screwed up my whole SVN because of a simple change in my domain DNS!

Their support is also a joke, cause they try everything to prevent you from contacting them directly. I really don’t know why I’m staying longer with them, must be that I’m just too lazy to transfer some of my old sites to another hoster.

In future I will transfer everything to Amazon, because there the administration is mostly under my very own control…

ReadBurner now filters Languages!

I have now implemented language filtering!
Though don’t expect it to be 100% accurate. It’s very experimental and works only good if there’s enough data.

Translation: the more items in one language are shared and the more people are sharing in that language the better the filter gets. Sounds strange, is strange 😉

Now I’m interested myself how this work…
If I it turns out to be a good-enough-solution, then filtering languages wasn’t too hard at all.

Oh. I forgot. I filter language that are similar and not every single language. If I made a big mistake with asian/persian or whatever language then please tell me.

Language filtering

I’m currently working on language filtering. I seem to have found a very easy and practiable solution. If it works out fine in my tests, then expect to have language filtering implemented in 1 to 2 days.

Multiple Items are a difficult thing. Can you help me?

Currently I’m counting multiple items based on links. The problem with links is that they are not so unique as I initially thought.

An example:

Lois Gray shared the item “ReadBurner Goes from 0 to 12,000+ In Four Days” which links to:

http://feeds.feedburner.com/~r/LouisgraycomLive/~3/
214788569/readburner-goes-from-0-to-12000-in-four.html .

I reshared the item using the most recent feed of ReadBurner but there the link is:

http://feeds.feedburner.com/~r/Readburner/MostRecent/~3/
214798652/readburner-goes-from-0-to-12000-in-four.html

The real link suddenly appears in the original:id field of the feed but I cannot use this field either, because sometimes google writes something completely usesless into this field.

Whatever.

I saw this problem coming from the first day, but due to the leak I had no time look more deeply into it. One thing I tried was counting by Headline, but it made some strange problems, so I threw the code away.

My questions to you:

• Can there happen something strange, when I’m comparing mulit-items by Title. Don’t know what it can be. Something like Charset problems in other lanugages or titles with html in it or what the heck ever?

• Is there maybe a much easier way to get a unique id of an shared item?

Let me know if you’ve got some ideas.

Who would have thought…

that the Arabic Persian Google Reader Community is so active?

Currently most items of the Popular Stream are Arabic or some Asian language. I love to deal with people from different cultures, but unfortunatly I don’t understand what the posts are about.

Neverthless, I think that over time the English linkblogs will have the greatest chunck. If not, I better start learning to understand those other languages 😉

RSS Feeds added

RSS feeds are finally availiable! You can find them on ReadBurner next to the respective tab.

Though I have to admit that they are kind of experimantal right now. My sortings are suboptimal. I have to change the DB a little bit. Gonna do this tomorrow.

Another thing is that I forgot to save the “content” field of the feed into the db so there are just headlines. I updated my parser but currently I haven’t updated the items. That means: old items have no content, new items (since 2 hours) have content.

I’m also working on user feeds but I ran into some problems. Have to change the db for this too, arghh…
Plz, give me feedback on the feeds.

Thanks Alex

Google Reader powerusers are smarter than digg.

Mike Reynolds to me over LinkedIn:

I think ReadBurner is great because Google Reader powerusers are smarter than digg.

That makes my day 🙂

New Linkblogs are coming in fast

I’m currently watching 239 linkblogs up from about 50 2 days ago.  That’s a lot of data to play with. It’s likely I’m adding a designer to this project to improve the look and feel.

A problem that’s popping up is different languages, especially those with no latin characters like Chinese or Arabic. I gotta find a practical solution for that…

RSS Feeds have priority

I have got tons of request for RSS Feeds. I’m gonna integrate them very soon. Has any of you an idea which RSS standard I should use? Or would you like a feedburner feed?