No, No, SolarSeek will live on. But not with me on board. I have quit the project last March. Read the details below.

We have planned to release an Alpha this Spring. We couldn’t hold the deadline. As a consequence I quit the project at the end of March.

I wasn’t really satisfied with the long release cycle for the next SolarSeek version eventhough it is an open source project and only one programmer works on it. Without giving the community some new code my motivation was very low to continue development.

This means there is a job opportunity for you. Do you like conceptualize user behaviours and needs, designing user interfaces, building them in Interface Builder or even programming them? Write an email to the development team.

I wish all the best to the whole SolarSeek team. My special thank goes to Erick Fejta who founded SolarSeek, to Marcelo Alves who currently programs SolarSeek and to Jonathan Kantrowitz who is Admistrator. They did and still do a great job. thanks for all the fun time.

auried

We have solved a huge memory issue in the new version of SolarSeek. Now we can move forward and release something very soon.

The memory issue was caused by some garbage sent by a peer just before closing the network socket. This could lead to a large message (MiBs) and we were allocating all required memory before reading the entire message. Now we’re allocating just 512bytes per message (this value is fine for almost all messages, except RoomList and SearchResults. The NSMutableData/SoulSeekMessage can grow up when necessary) and reusing some NSMutableData very often, avoiding stressing the Garbage Collector.

I’m redesigning some of the SolarSeek icons to make them look better or more accurately representing their functionality. Here is a preview of some of the new icons. The first row shows the preferences toolbar, second is the main window toolbar and the last row is the sidebar. It’s up to you to find out what each icon does exactly ;-).

The current SoulSeek protocol doesn’t support multiple sources downloads. This means you have to wait in a queue till you are first and your download starts. For british people this wouldn’t be much of a problem but for the rest of the world to queue means to loose time.

In real life queues are a mysterious thing: on one hand they can be huge but you only wait five minutes but on the other hand five persons can make you wait an hour. SoulSeek is not different in this manner. You see a queue number but you don’t know how long you have to wait. Maybe you observed yourself already comparing upload-speed with the queue to get some idea how long you will have to wait. That’s all right if you like maths but for the rest of us the situation isn’t satisfying.

That’s why SolarSeek introduces the “dispatch” semantics. It’s a time indication that shows you how long it will take until your file will be dispatched or in other words downloaded. This means you don’t have to interpret an abstract queue number.

dispatch time = (( queuelength / uploadslots ) * ( totalfilesize / totalfiles )) / uploadrate

In addition a result with no queue will be rated to indicate which one has best quality at the highest speed. And all this again without concentrating on number that you would have to interpret.

dispatch rating = ( mp3rate / ( totalmp3rate / totalresults )) / ( uploadrate / ( totaluploadrate / totalresults ))

What do you think about the dispatch semantics?

Update: We probably call the column “Queue Time” instead of “Dispatch”.

After Leopard was released, and Iwan, Marcelo, and I got our hands on it, as a team we decided to make SolarSeek Leopard only. This decision came after several improvements with Objective-C 2.0, Interface Builder, Core-Data, and a few other components. There have been no commits into the main trunk, as we have been making all new commits to the new Leopard branch. For those interested, it can be found here.