Learn my Linux Bootcamp…all from within a web browser window

I awoke yesterday to see a lot of twitter notifications on my phone. Sometimes this happens when I've written a post on this blog, but I hadn't added anything for over a week. Turns out that the activity was triggered by this tweet by Richard Smith-Unna (@blahah404 on twitter):

As the screenshot below indicates, Richard has worked some amazing black magic to enable a single browser window to contain a fully interactive terminal as well as a file viewer/navigator; all alongside a (slightly modified) version of my original Linux bootcamp material.

Click to enlarge

This new interactive command-line bootcamp is a wonderful resource and means that the only barrier to learning some simple, but powerful, Linux/Unix commands is the availability of a web browser.

Richard explains a little about how he put all of this together:

The Infrastructure, including adventure-time and docker-browser-server, was built by @maxogden and @mafintosh. The setup of this app was based on the get-dat adventure.

Slides from my exit seminar

This morning I gave my last presentation at UC Davis. My highly informal exit seminar was a great opportunity to reflect on some of the many projects I've been involved with over the last decade here at Davis. Thank you to all who came, and a special thanks to Ian Korf for his kind introduction.

I include the slides below, but note that some of these slides won't make much sense without the narration (and you also get to miss out on two embedded videos). There was some video recorded via the Periscope app, but I found out today that Periscope only keeps video around for 24 hours, so unfortunately if you didn't watch the video when you had the chance it is now lost.


2015-11-21 12.34: Updated to reflect that Periscope video content is no longer available.

JABBA vs Jabba: when is software not really software?

It was only a matter of time I guess. Today I was alerted to a new publication by Simon Cockell (@sjcockell), it's a book chapter titled:

From the abstract:

Recently, a new hybrid error correcting method has been proposed, where the second generation data is first assembled into a de Bruijn graph, on which the long reads are then aligned. In this context we present Jabba, a hybrid method to correct long third generation reads by mapping them on a corrected de Bruijn graph that was constructed from second generation data

Now as far as I can tell, this Jabba is not an acronym, so we safely avoid the issue of presenting a JABBA award for Jabba. However, one might argue that naming any bioinformatics software 'Jabba' is going to present some problems because this is what happens when you search Google for 'Jabba bioinformatics'.

There is a bigger issue with this paper that I'd like to address though. It is extremely disappointing to read a software bioinformatics paper in the year 2015 and not find any explicit link to the software. The publication includes a link to http://www.ibcn.intec.ugent.be, but only as part of the author details. This web page is for the Internet Based Communication Networks and Services research group at the University of Gent. The page contains no mention of Jabba, nor does their 'Facilities and Tools' page, nor does searching their site for Jabba.

Initially I wondered if this is paper is more about the algorithm behind Jabba (equations are provided) and not about an actual software implementation. However, the paper includes results from their Jabba tool in comparison to another piece of software (LoRDEC) and includes details of CPU time and memory requirements. This suggests that the Jabba software exists somewhere.

To me this is an example of 'closed science' and represents a failure of whoever reviewed this article. I will email the authors to find out if the software exists anywhere…it's a crazy idea but maybe they might be interested if people could, you know, use their software.

Update 2015-11-20: I heard back from the authors…the Jabba software is on GitHub.