Tuesday, 26 August 2014

Post post doc

Continuing what has become a quarterly theme here is summer's post:

One reason for a drop in blogging productivity is that I have been severely distracted by that thing that many postdocs are -- career. It's a pretty neat job being a postdoc. Like a PhD you've got more time to dig deep into a topic but you're more on top of your skills. It's flexible, it's a great way to get work abroad, you can wear a T-shirt to work and so on. It's a nice job.

Eventually you do start to want a permanent position. Now in academia these are hard to get and there's very little geographic control. Many have written long and angrily about this, that's not what I want to get into. It is what it is. From a practical point of view it as an excellent time to think about whether you really want that academic job -- do you want to stick of twist? In my case I've gone for twist.

This autumn I'll be making the move from statistical physics to statistical, er, statistics. My physics work was gradually slipping towards data science and I can't resist any longer (it's the sexiest job of the 21st century don't you know!). Actually I want to post about that physics/data work at some point but need to wait for reviews etc.

I'm really excited about my new position and all the new things I'll be learning. If possible I'd like to keep this blog going, I think there's a decent cross over with what I'll be doing next. The themes might switch more towards data than soft matter but statistics are statistics. It'll be a little sad leaving physics but I'll always see the world through a stat-mech lens, which is no bad thing for anyone in my opinion.

Thursday, 15 May 2014

Journals for e-readers

One thing that makes me cross is that despite the terrifying amount of money our library pays to buy back our research in the form of journals, they're still not terribly easy to read. I've got an e-reader now and I'd like to read things on that, just the sort of value-added that the publishers could do. Unfortunately everything is still just a pdf file only to be printed on A4.

There are some utilities for coping with this but it's not really ideal.

I wanted to see how tough it is. So I tried to convert my last paper into something that would look nicer on an e-reader (in my case a kindle). The paper was written for an APS journal using the REVTeX 4.1 package. This makes it very easy to write papers formatted for APS (and possibly some others as well). The answer is that this was the best I could get it using REVTeX.

which originally looked like this:
or here if you have access

It's actually not too bad! The abstract's gone a little wrong and the font is not strong enough, but it's not a disaster. It proves to me that the guys who make REVTeX could quite easily make a beautiful e-reader mode full of useful options. I made the citations clickable links for example.

To get this working I simply replaced this line
\documentclass[aps, prl, twocolumn,superscriptaddress,amsmath,amssymb,floatfix]{revtex4-1}

with these lines


Without the second one it doesn't seem to work very well. If you've got a better (and just as easy method) then leave a comment. I could have made it better by dropping REVTeX and customising every detail. Frankly that was proving a lot of work and that's not what I believe LaTeX should be about.

From now on I'll be uploading an e-reader version as well as an A4 version for my papers.

Royal Society of Chemistry, with a bit of fiddling looks fantastic on the kindle:
Had to dig about a bit to get this one-column (there's a \twocolumn[ half way down the page in the template that needs removing). Fiddled with the margins and widths a lot. See the .tex file here. Might go back to the APS paper and reduce the paper size even more. This seems to work quite nicely:



Got the RSC send-to-kindle button working and this just sends the two-column pdf. I guess the best hope is converting the html version then.

I've found the best way I think now. It's to skip pdf altogether and go via html. Using htlatex I compiled the same original latex file into this webpage:
It's a bit broken here and there (you must use revtex4 and not revtex4-1, one day I'll go back through and work out all the settings needed to properly convert a revtex made file. Mostly it works.

Next step was to download the KindleGen utility and use that to convert the html to a .mobi file format. This is what I then put onto my kindle. If you download this and put it onto your kindle
you'll see that this is basically perfect for what you want from a Kindle version of a paper. For sure some things need fixing, for example the equations are a bit small. I'll work all of that out and make a separate post.

The bottom line for this whole thing is that if I, a mostly lazy man, can get half way decent conversions of my LaTeX files onto an e-reader in an evening, then the journals could easily do this.