thelittlelisper

I'm very happy with xmonad. I'm running a very minimalistic desktop consisting of fast applications that do one thing, do it well, and can be controlled from the keyboard (zsh, urxvt, emacs, remind, mutt, zathura...).

As a very welcome side-effect, the whole thing is very light. It runs in a few tens of MB.

It is also compulsory if you follow the International Baccalaureate program, under the name of Theory of Knowledge. I think this is extremely useful for future social / natural scientists and engineers.

org-mode has a pretty good phone app for iPhone & Android for when you're on the road: http://orgmode.org/manual/MobileOrg.html

Well, Janellia is actually in my research area, so you pretty much nailed it.

The major problem with Academia (actually the Academia subset where I'm unfortunately stuck in) is entirely different.

Here it is very rare that people above grad students and perhaps postdocs do any research at all. They're just managers, except in rare occasions or in the very best places. Naturally, it is extremely difficult that anything good science-wise comes out of this.

The structure is frighteningly similar to a corporation, except that corporations usually end up dying if they're inefficient, whereas research institutions only need to produce papers to get funding—and that's an easy game to play.

I can hardly think of Professors in say a Math dept. managing their students. Yes, they're advisors, but they also conduct their own research.

This would be brilliant. Does any other framework support this? (outside Clojure)

Thanks. Care to elaborate on this?

It's difficult to leak details and make an impact. Academia is really screwed up. 90% of what gets published is a blatant lie, but nobody seems to care. Most researchers live in a distorted reality.

I wholeheartedly agree with this. I work at one of the top cancer research labs in the world, and you'd be surprised how dysfunctional our organization has become. It's perfectly described by what edw519 has outlined.

Seeing millions of dollars poured in projects that go nowhere since they are executed by incompetent workers lead by machiavellian managers is horribly depressing.

I feel that a startup setting would be incredibly more efficient to do what we do. But of course it'd be difficult to fund it, unless it's operated as a charity or it's backed up by the gov't.

In the meanwhile I'm acting as a stoic and enjoying some tiny victories.

As a zsh user, fish looks fantastic thanks to its lightness, responsiveness and default settings.

I'm only slightly unhappy about fish requiring python (actually python2 in Arch Linux). AFAIK, this is needed for the web server configuration interface. It'd be much better if fish could be built with this as an addon or optional dependency.

Well, running a lightweight desktop helps. I cannot say how much, though. Try monitoring wakeups using powertop while your machine is iddling. Anything above 30 or 40 is bad.

If your connection drops often it's a nightmare.

I only recently discovered isync (mbsync), which is sad given that it is crazy fast and written by the mutt author.

offlineimap is cool, but it is sadly quite slow and crashes often. New releases usually bring as many bugs as they fix.

Mutt / Notmuch + isync (mbsync) / offlineimap works fine...

I always go back to mutt. It does one thing, read email, and it does it well.

The key is to set it up appropriately. Don't use mutt for downloading your email, especially if you wanna deal with IMAP. Use isync (mbsync) for that purpose (I used to use offlineimap, but it's buggy and not very fast).

There's also notmuch, which I use for indexing my email. It also contains a pretty good email reader.

These are my energy tweaks for a 2008 MacBook, I don't have my x220 here, although they are machine independent:

echo 5 > /proc/sys/vm/laptop_mode echo 1500 > /proc/sys/vm/dirty_writeback_centisecs echo 0 > /proc/sys/kernel/nmi_watchdog

echo 1 > /sys/module/snd_hda_intel/parameters/power_save for i in /sys/bus//devices//power/control; do echo auto > $i; done for i in /sys/class/scsi_host//link_power_management_policy; do echo min_power > $i; done

for i in /sys/class/rfkill//soft; do echo 1 > $i; done

For the kernel parameters, a simple Google search should get you going.

What's the watt consumption you achieve? I must note I run Arch Linux with Xmonad and no desktop environment.

Just 3 things:

- Install powertop and enable all suggested tweaks

- Depending on your kernel you should downclock your i915 graphics card and force pcie_aspm

- Later kernels (>3.6?) have a power regression which leads to higher consumption in Sandy Intel architectures

I'm not saying I would. I prefer the trackpoint. But they should provide a decent trackpad or get rid of it altogether. I don't need both. This would make more space for palmrest, which is too small in x220.

Have you tried diagnosing with powertop?

It runs fine (i haven't tried many multitouch gestures).

Most hardware is Intel (except wireless card, which is a b43, also supported).

With a few simple tweaks I can achieve equal or better battery consumption than OS X.

My favorite options are the Air (11 / 13) and Thinkpad x220.

Other options include Thinkpad x230 (newer than the x220) but worse keyboard, x1 Carbon, Dell XPS, Chromebook Pixel, Asus and Samsung.

I ended up buying both an Air 11 (my own money) and a x220 (my employer). Both machines are really well supported under Linux (most hardware is Intel). The only drawback is their low resolution screen, which IMHO is not a dealbreaker given how small they are.

I manage to achieve 5-6W in both, which leads to stunning battery peformance. This requires some powertop monitoring and some simple tweaks.

x220 is great because it is serviceable, has an IPS screen, a full keyboard, a non-ULV processor, better connectivity, and a more sturdy design.

The Air has a much better touchpad, is more silent, and comes with a blazing fast SSD from Toshiba (64 or 128 versions).