BikeIM 0.5.1


May 5, 2017

On Thursday of this week (May 4, 2017) I released a mail client. It's pretty humble, but I've been using it for a decade and so it was worth it to me to iron out a handful of bugs and make it good enough for other people to use. Currently it won't be much use to you if you run Windows, Mac OSX, don't run your own mail server, or don't understand what fetchmail is*. But those things are fixable and I intend to fix them in the months to come. That means this blog post doesn't have to cater to end users, so I won't attempt to. This blog post will be about how I came to write a mail client and why it makes sense in 2017.

* The wording of that sentence was pretty bad so I'll reverse it. BikeIM-0.5 is usable to you if you run Postfix, if you run Dovecot, if you run OpenSMTPD, if you run fetchmail, if you use GnuPG, if you use Mutt, if you run Linux or *BSD, if you use Git, and understand how to report bugs.

Read more »

AltSci Kanji Classifier


Mar 21, 2017

learn_kanji-0.4.tar.xz [sig]

This blog is a continuation of this blog post which is a pretty good mix of technical and creative projects that this blog was originally written for. If you look at some of the way-way-back posts you'll see some of the things I've thought about and tried to work on from this perspective.

Read more »

AltSci Concepts in Python


Mar 1, 2017

It took quite a few months to get the time to work on this blog. I had AltSci Cell and AltSci Blog updated to the new blog system back in November. I ported most of AltSci Concepts to the new blog system a few months ago, but I didn't want to convert it until I had comment working. You can see there are no comments at the bottom of this page because I'm not ready to backport the comments to AltSci Cell yet. But if you go to any of the posts on AltSci Concepts you'll be able to post your comment.

In all the years of AltSci, only a handful of people have commented. The main purpose has been to collect a large corpus of spam (389245 spam comments from 67085 unique IPs, 9683389 csrf spam attempts rejected) which I've had to filter at times. But the opportunity to post responses is a valuable part of the Internet. That's probably why sites like Hacker News and Reddit are so valuable: they let you comment on important topics where the comments are disabled or too toxic to post in. That doesn't mean that these sites are infallible. In fact, that's a lot of the controversy in any open forum -- people post in a thread that is going to stay linked to that website for a long time and time increases a search engine rankings. When should the post be deleted? When should the admins leave it alone? These problems are not trivial to solve. If you think they are... please create a thread to discuss this post and e-mail me a link.

Read more »

Natural Language in Small Wide World


Sept 1, 2016

Yesterday I published a small piece of software to Small Wide World's git to very little fanfare. It was a generalization of a bad piece of software I wrote the day before. It uses NLTK to perform a simple task: parse a simple sentence which follows the form "subject verb object" with optional additional information starting with "because". Examples of this grammar include:

GnuPG is software
IRC is a protocol
software implements a protocol
Javantea is human
AI3 is software
Javantea wrote AI3
Javantea writes software
Javantea writes English
Javantea reads German
Javantea reads Japanese
Javantea reads Portuguese
Javantea reads Spanish

nlp1.py creates this graph of the relationships:

Read more »

« previous next »