Reaktor User Library vs. Processing

This may crash your browser, steal your credit cards and possibly impregnate women of age in your household, but, over the weekend I made this as a goof.

The code isn't optimized, the logic for creating node paths is beyond terrible (if your search is too big it'll run out of memory) and it's about a 10 meg download, but it should run in any java capable browser.

http://www.remix.net/static/applet/

Try searching for 'FFT' in the description field.

 

Looks really cool. The text is just a touch too tine to read on my screen. Sort of unclear what the different branches means...

 

sorry - no (SAFARI)

ciao herw

 

Very buggy indeed
I tried to open it with firefox and osx10.4, at a certain point I saw menus like in the user library main page. Is it a new version of that you made with Processing? I like that program: it sends osc, so it could be the future for GUIs, also with Reaktor...?

 

That's interesting- all the variables I tried came up with the associated uploads listed. An interesting search tool, to say the least...

ew

 

Yeah, that's kind of a tough problem to be honest -- basically each branch represents a series of parent/child relationships -- so any time you see a circle that's a parent category node, but labeling them ends up being really ugly. I'm considering a few ways of dealing with that.

And the text is super tiny simply because large results end up being a shotgun of text in the face. Kinda thinking the next phases should be to put a result limiter or a timeframe window so you can pair them down.

After I get those done I'll work on getting 'click on a library entry node, get the details'..

Processing is pretty nifty.

I will say this, any suggestions on how to move forward with this tool will be heard!

 

Forgot to say I really welcome this initiative! There are soo many hidden gems in the UL.

First, I'm not sure what parent - child means in this context? Is it a referal by in the text?

One way of finding things in a huge library is by tagging. last.fm e.g. uses that to some success I'd say. So if users could put tags on ensembles it could grow over time.

Combined with favorit lists and if-you-like-this-you-might-like-that lists it would be a real power UL mining tool.

 

Everything above a library entry is a category. So, there's "Synthesizers >> Experimental" etc, so those get broken up into nodes so that the resulting structure ends up being

Synthesizer->Experimental->SomeExperimentalSynthesizer.ens

Tags are do-able, that might be interesting to combine with text searching.

I rewrote all the node code so it's real OO now, the particle system is kind of slow, but.. Meh. I suppose in theory the rating system is a 'you might like this' indicator but.. I never found it to be that useful.

If NI had an API to interface with, the tool could really be nice, but.. It's basically stand-alone and a pet project so I doubt it'd 'really' take off. But.. it is much more zippy than the existing search iddinit..

-n