Wednesday, January 28, 2015

Strange hammer

When I first saw these parts I originally thought they were junk.

Then I figured out they slipped together to make a hammer. This was kind of interesting.

The outside of the head seems to be hand forged and heavy steel, but the face of the hammer was just soft wood and has been beaten in over the years. I wasn't sure why anyone would make such a thing or what they were hitting with it.

Then it dawned on me, this is the perfect hammer to hit the back of a chisel with. The heavy metal provides mass and the soft wood in a concave shape is gentle on the back of the chisel and won't deflect. Whoever made this was a genius.

I fixed the cracked handle by pushing wood glue into the cracks, putting some packing tape on a clamp and holding it all together until it dried. It ended up working perfectly. The handle is back to it's original strength and is now usable with it's perfect fit and original patina.

The rust on the head had to go. Die grinder with some scotch-brite made short work of the rust and shined up the metal.

I coated everything with a bit of boiled linseed oil so it won't rust again. I've found the trick to applying boiled linseed is to apply generously but then do your best to wipe it off with paper towels. If you don't get aggressive wiping it off you'll end up with a stick mess on the metal when all you really need is a very fine layer to protect it.

Friday, January 23, 2015


New project.

Velociraptor means "swift grab" and this software is a parallel tool that uses your AWS EC2 account to swiftly fetch lots of web content.

Velociraptor takes a list of urls and launches parallel AWS EC2 instances which then run wget in parallel to fetch the html and images into Web Archive .warc files and store them in an S3 Bucket. Parallelism is managed by GNU Parallel and AWS is managed by AWS CLI. This project is not associated with AWS and may use other cloud service providers in the future.

Thursday, January 22, 2015

Data Science tactical and strategic

Selling data science at tactical level is a matter of patience, you have to convince people to pause and use data before making tactical decisions.

Selling data science at the strategic level is a matter of trust., often lots of thought and time goes into strategic decisions, you have to convince people the analysis of the data is correct.

Tuesday, January 20, 2015

Shoe rack

I make a habit of snatching the pallets when we order servers rather than letting them be thrown in the trash. Colleagues at work have taken note of this and started setting them aside for me, so I now have more than I could ever use. This is fine, my wife and I sometimes like rustic looking things which also seem to be in style right now. What doesn't get turned into projects makes great wood for the firepit in the back yard.

This weekend I made a shoe rack for the bottom of our closet.

Here's a picture of the original problem, unsorted shoes.

After pulling nails and other metal out of the pallets, I cut them to shape on the table saw. Then glued and used my pneumatic brad nailer to stick them together like this.

I didn't use any finish on them. Just left as raw wood. The bottom of the closet is a stable and dry environment, the plain wood will literally last for centuries like this.

I screwed the runners to the inside of the closet wall and then just set the shelves on them. This way, if needed, we can just lift the shelves out without having to undo any screws.

Here's a picture of the finished project.