Turbo Slug

Well I finally got around to de-underclocking my NSLU2 (aka Slug).  It took about 10 minutes start to finish, so I really have no excuse for why it took me this long to get around to it.

The NSLU2 Linux wiki has very detailed directions on cracking the case open.  Go slowly to avoid snapping off the plastic tabs.

Now comes the fun part, modifying the hardware.  It turns out that for whatever reason the NSLU2 was underclocked, the wikipedia article discusses this and other interesting facts about this device.  While this is a discontinued device, they are readily available from amazon still.  All we need to do to allow the Slug to run at its native 266MHz is remove a resistor.  I used a pair of tweezers.

If you look closely, you’ll see the tiny resistor in the center of the palm of my hand.  I later dropped it onto the floor and lost sight of it.

Putting the unit back together isn’t documented as well on the wiki – I found that it was easier to insert the circuit board into the light grey “front” and then snap the back on.  Again, go slowly to avoid breaking any tabs off.

Having installed a version of Debian on it the other day, I was able to ssh into my new “turbo slug” and cat /proc/cpuinfo which showed me BogoMIPS: 266.24.  Success!  It does seem to be more responsive.  I’m tempted to try tinkering with doing a memory upgrade, but that may be beyond my soldering skill level.

Playlists

It seems some people are really into creating playlists.  I’m not one of them.  While I have a lot of nasty things to say about iTunes, the smart playlists feature has been the only way I’ve bothered to make playlists.  At least until now.

The other day I was thinking that while its nice to have iTunes recommend songs that you might want to buy that are similar to what you’re browsing in your music collection, and Amazon has lots of recommendations on what you might buy based on your previous purchases.  No one seemed to be harnessing the relationship between things that I already own.

Well, it turns out there is a bunch of stuff out there that does exactly that.

At first I ended up at wikipedia on music recommendation systems.  Further digging turned up something called The Filter which was blogged by Wired a while back, unfortunately it doesn’t seem quite ready for prime-time so I keep hunting around.   Then I came across MusicIP mixer, and a related comparison review article on lastfm.

MusicIP seemed like the best, most available candidate, so I gave it a shot.  The free version is limited, but not crippled.  It installs as a stand alone application, but is aware of iTunes and will read settings from there to figure out where (and what) your music collection is.  Out of the box, all you do is pick a song (or a few) and ask to make a mix.

I picked a random Haujobb track and asked for a mix of 75 songs (the limit for the free version).  The sample_mix was a fairly good selection of the music in my collection that I’d consider to be fairly compatible.  There were a few odd songs in the list that I trimmed out by indicating “less of this” and it shuffled things around (and I suspect remembered the error for later).

I then burned a MP3 CD for the car with these 75 songs (some 500MB).  As an aside, I’m terrible for updating my driving tunes, so what I had in the car was pretty stale.  Well, its been 2 days so far and I’m very pleased with the experience.  This is a much better mix set than the smart playlist I primarily use on my iPod.

One note about the MusicIP system, it needs a lot of processing time until it has consumed your music collection.  I ran my PC for well over 24hrs processing 3500 songs.  This is a one time hit, and I consider it a fair tax for the free version of the software – especially if it gives me an easy way to make new mix CDs, or new playlists for my iPod.

Be it through some social network, algorithm, or mixture of the two – I hope that this type of technology grows in popularity.  As we gather more and more media, we’re going to need better ways to interact with it – and insisting that we manually tag, sift and sort our data simply won’t scale.

Imagine if your email could be grouped and sifted based on how others had handled the same messages?  How about if your PVR could just record the same shows your friends are watching?  If you always answer “yes to all” on that dialog box, why not let me make it a sticky setting?

WordPress Exploit

I recently upgraded to WordPress 2.5 – and in the process of doing so, I noticed something funky with my older 2.3.3 installs claiming to be 2.5 already. I thought it was odd – but didn’t immediately come across anyone having reported strangeness here and so I just ignored it.

Now that 2.5.1 is out, I thought I’d go upgrade again. Well, after the upgrade I was still having the dashboard tell me that I needed to upgrade. Odd. This time a web search did uncover information that was relevant.

Details on the wp-info.txt exploit are interesting. It seems to me that several problems are being lumped into the one discussion, but I found some helpful advice to help clean things up from the links provided there.

Symptoms:

  • Presence of wp-info.txt
  • Displayed version changed without upgrading.
  • Database modifications
  • New files ending in _new, _old, .pngg, .jpgg, .giff appearing inside writable directory

Continue reading “WordPress Exploit”