code

In search of a prefix: "//" should indicate a URL

Schuyler — Sun, 26 Jun 2011 23:22:36 -0400

You know it's an email address when something's in the form foo@example.com -- the '@' and the domain gives it away. Nowadays people know that @schuyler1d probably indicates a Twitter account by the '@' prefix, and #urlprefix is a subject tag.

But the most common Internet URI of all still has an unwieldy "http://" prefix or "www." too long. We used to be able to tell, because everything ended in ".com" (even when it didn't) but now URL shorteners and cool domains have hacked all the country codes to bring us delicio.us (which had to register delicious.com, because no one could tell that it was a URL), bit.ly (using Libya's registrar). And now ICANN is allowing a glut of TLDs (the ".com" part), even though no one uses them.... because everyone expects to type ".com"

Now Mozilla and Google are starting a trend to remove the icky prefixes from the web browser -- maybe removing the URL bar altogether, because the "http(s)://www." is ugly and ultimately confusing to users who shouldn't have to care about protocols (in the near future, https or another secure protocol should be the default).

This is going too far, because we still need a way to indicate that we're referring to a web address in text, including in the address/title/url/search/awesome browser bar. It's a larger issue that extends to print -- business cards, advertisements, tweets -- they all need a bunch of extra characters to communicate to people that YOU ARE LOOKING AT A WEB ADDRESS.

Let's stop this insanity. Twitter users (and $larry_wall) showed us the way. "//example.com" should mean you're looking at a URL. This site is //skyb.us. This'll look a whole lot sexier (and more concise) in ads. Google can use the spdy protocol without lying to its users, and maybe airlines can start using the //wright.aero TLD.

Distributed Version Control for Freedom

Schuyler — Sat, 09 Jan 2010 13:59:09 -0500

Short technical note: I've been using Mercurial for my personal projects, but in succumbing to community momentum (and really, a much better submodule/svn:externals story) I'm switching over to Git. For some reason hg2git converters are many and generally broken. I tried a bunch, but few of them built/worked. The one that finally did is Antono's hg2git.

Distributed version control is one of these secret stories going on in the land of programmers that needs to get out into the imagination of the population.

In short, your next 'wiki' for public consumption and benefit, should maybe be hosted at GitHub.

Let's take a good example of a public document which deserves scrutiny and every party would like some changes to it: US Federal Law. Now, there's a process to change the canonical version of this document. It involves a lot of elected officials practicing strange rituals in public and private, and then suddenly there's a different document which the US courts and executive branches recognize as the current document. You can download it from the Intarwebs.

However, what changes is the ACLU keen to change? They write a bunch of public position papers and their lobbyists try to work with lawmakers for certain language. But wouldn't it be cool, to see the EXACT CHANGES that the ACLU would make if they ruled? Maybe it would be boring, but if every lobbying/political group did this, then we might have a pretty cool patchwork of legal frameworks to choose and examine as citizens.

Back to those strange rituals that elected officials do to make the law (and the SCOTUS which is not elected)--the reason this is less interesting than it could be is that the rituals are so strange and non-sensical that the source of law couldn't possibly come from one source. What if we could track WHERE the laws came from? Then we could compare how much Walt Disney Corporation gets to write law compared with PETA or corn farmers.

We'd have a much better understanding of how the US legal system came to be, and citizen-based lobbying groups could rate representatives, not just on votes, but how much law got in to the rules of the land.

Unfortunately, we don't have anything like that. It would be nice if even the Congress did this internally, but....that would make sense.

Fortunately, free computer-dom is different. The basic laws which govern free computers everywhere required just such a system, and that law is called Linux (for an easy start go to Ubuntu). From that first link you can trace the history and who did what and where they worked, etc.

Now just because this is most easily understood as useful for a big document like US Law or the Linux Operating System, doesn't mean it's not useful on a microscale. What about Subletting contracts, or other legal contracts. Mostly we get our lawyers to write them or download an example, and then we change them a bit. The same analysis of who wrote different contracts and details about how well they fared in court and in practice might also be useful. What about seeing the same thing for a news article, as a story develops--what new sources become useful, etc.

How is this different from Wikipedia? Most of it is the same. Wikipedia is based on 30-year-old technology that programmers used to track different versions of software and brought it to the web for encyclopedic topics. The difference is in the decentralization. Why does Conservapedia and Deletionpedia and others have to start from scratch? Once their viewpoint on a particular article becomes useful, then there should be a way to automate bringing in the work done on one site into another. And THAT's what Distributed Version Control is about.

Non-linear Writing Systems: Radical Culture Hack

Schuyler — Sat, 05 Dec 2009 11:01:55 -0500

Have you seen the US military's tangled web of a plan for Afghanistan? I really hope they paid that group to make it look complicated for PR purposes, rather than it being something they actually rely on. That is a BAD visualization, which could be much better organized.

As it happens, I have a radical culture-hack qua artistic-project that could theoretically help with problems like this some day. My website is out-of-date and it's complicated, so I'm mostly shy about it. As is typical, the Internet eventually showed me that I was neither the only nor the first person with my idea. Someone asked me recently to write up a short summary about the idea for a tech audience. Since they didn't use it, I figure I can put it here:

It's something of a coincidence that we still write in continuous unbifurcating lines. The first writing was designed for accounting. Scribes would trace out columns and rows like graph paper before filling in the boxes with words. If you've seen a printing press' block-type, you'll know we haven't evolved much further since then.

Glide by Diana Slattery

The original World Wide Web gave us hyperlinking; an improvement, but this is like running around a first person shooter without a map--perhaps an enjoyable waste of time, but a slow way to get to your desired destination.

Pinuyo by René Uittenbogaard

Where are the flow-charts for learning about things other than a state machine? We have to make them! And yet, whereas flow-charts provide a proof that we can build them and even aid our understanding, it is the map metaphor that makes Non-Linear Writing Systems (NLWS) deeply compelling. Navigating a legal system or esoteric philosophy can be made as simple as navigating a city walking-tour.

Ouwi complete first chapter
of Tao Te Ching by me!

Whole fields of thought could be mapped as a fractal-like text book. In poetry and narrative, people's journeys could intertwine on the page as they do in the world.

Making a NLWS with fractal-like properties is both fascinating and a significant challenge. For those with a mathematical bent, there are many useful insights in planar graph theory. Anyone can start, though, by coming up with creative ways to organize and summarize a story visually.

Why to turn your bookshelf into a hard drive

Schuyler — Sun, 08 Nov 2009 11:41:38 -0500

This post is the beginning of a series on how to destroy all your books while preserving all the information inside them (yes, even your margin notes), for a better and more portable reading experience. Before I go into the how, here is the why...

My wife and I are book fiends. By that I mean we lost the war a long time ago. Several bookshelves are double-layered and stuffed to the brim and book piles teem from behind our couch. We are good customers to the publishing industry. Despite working in a library, I still buy most of the books I read.

I happen to own a 2lb laptop which I can carry everywhere, and love it. On the other hand, I'm still working my way through Against the Day after a year or two, and I mostly blame the fact that it's too heavy to bring on the subway conveniently.

A lot of people complain about reading from screens, but it's getting easier every day. I don't like the DRM on the Kindle, but e-Paper in better forms is coming as soon as this Christmas.

Speaking of DRM, I buy ebooks, too, when they're available--I'll even pay a premium (despite how ridiculous that is, considering the decreased cost for the publishers). They're only sold DRM-ified, but there are some tools to fix that.

That shores up the problem for some new books, but not all of them (I can't find Against the Day in .mobi format). It also doesn't solve the problem that the books are already taking over my couch.

Thus, we come to book scanning.

Book Scanning

First, you might say Google is doing the work for us. They certainly have scanned a lot of books, but unless you only read books from before 1923 or the US Congress suddenly gets a clue, most of your bookshelf will be unavailable that way.

Even as old an opus as Shakespeare really benefits from contemporary editing and analysis.

If you want to keep your books, and have a lot of time to turn the pages, then check out the BookLiberator Project. I don't have that kind of time, nor do I trust my home electronic-craft-project skills.

For me, the whole point is really to get rid of my books while preserving my margin-notes and have my entire library available on my laptop (and backed up to insure against a fire or flood). So, I bought a scanner with a friend and the journey has begun. My bookshelf is emptying a couple inches at a time. Future posts will explain in detail how we're going about it.

Switching to XMonad

Schuyler — Tue, 03 Mar 2009 17:34:45 -0500

I'm learning that the more work habits I adopt from a particular coworker and friend, the more efficient I am (aside: The last one I adopted was a monthly subscription to Odwalla bars at work). I've looked lustfully over at his xmonad desktop more than once before. The basic idea is:

I should NEVER see my desktop background--I want the window manager to configure the window sizes so everything is optimized for maximum space
Using the mouse to move between windows and tasks sucks! I want to use my keyboard

I'd given up after a couple dives, due to the learning curve--not the learning curve of the environment, but the learning curve necessary to configure it correctly in order to be efficient. This time, after a couple of labyrinthine Carnot paths that ended up being Work=0, I've come up with a much shorter transition plan for those interested in taking a plunge.

So, assumptions are: your average Ubuntu Gnome desktop environment. I have two monitors, so I talk about that case, but having one or three has the same instructions.

$ sudo aptitude install xmonad libghc6-xmonad-dev libghc6-xmonad-doc scrot

Install your xmonad config script.

This is the scariest step really. You can look at all your options later, but I found the current defaults mostly work.
The damn thing is in Haskell, and it seems they've refactored the API several times, so a lot of sections of published config scripts on the web will NOT work.
Start out with mine, which should get you started, anyway.

import XMonad
import XMonad.Hooks.ManageDocks
import XMonad.Hooks.EwmhDesktops
import qualified XMonad.StackSet
import XMonad.Util.EZConfig
 
main = xmonad $ defaultConfig
                { manageHook = manageDocks <+> manageHook defaultConfig
                , logHook    = ewmhDesktopsLogHook
                , layoutHook = ewmhDesktopsLayout $ avoidStruts
                                 $ layoutHook defaultConfig,
          modMask = mod4Mask
        
                }
    
        `additionalKeysP`
        [("M-", windows XMonad.StackSet.swapUp)
                ,("M-", windows XMonad.StackSet.swapDown)
            , ("M-f", spawn "firefox")
            , ("M-o", spawn "emacs")
            ]

Two things you might want to change:

Note that I have my text-editor, Emacs as Mod-o. You might want to change this to gedit or vi, or whatever
Later, you can add other keyboard shortcuts (or even sequences, like 'M-x f')

Put that config file in ~/.xmonad/xmonad.hs.
- If you start hacking it, you can run $ ghci ~/.xmonad/xmonad.hs to make sure it's valid.
Now edit/create a file ~/.gnomerc with the single line export WINDOW_MANAGER=xmonad.
- NOTE: If you give up, or don't like XMonad, just remove that line, and logout/login again!
If you have two monitors, one thing that I did more easily before the switch was adding a new GPanel to the top of the second screen, and adding a second Workspace Switcher widget, so both screens have one. This widget is non-ideal for how xmonad thinks of screens, but until something better comes along....
OK, you're ready to try XMonad out! Just logout and log back in (or restart, if you want to wait an extra minute).

Getting around

Print out this cheat sheet
Read the Guided Tour, and if you have two monitors, The idea of how screens and workspaces relate is important. I think the rough order of important shortcuts (the Windows key is the 'Mod' or 'M' key here, btw):
- Where is your attention: M-w (left monitor), M-e (right monitor); M-j, M-k (next/prev app on monitor)
- My current focus needs more/less space: M-<space> (cycle through different window configurations); M-h, M-l (increase/decrease window/app size);
- Switch workspace of current monitor to: M-[1-9]
- Send current window/app to workspace #: M-<shift>-[1-9]
- Give me a new.... M-<shift>-<Enter> (terminal) M-o (editor); M-f (browser)
- Move current window/app to different positions in current layout: M-<Up>, M-<Down>
Figure out how you're going to organize your desktop. I'm still working on this, but the main one is that Email and IM contact-list is on workspace 5. IM talk is on workspace 1, with terminal.

Dear Mozilla

Schuyler — Tue, 03 Mar 2009 16:10:13 -0500

Please pay attention to my adopted bug, with test cases. If you do, I'll eventually be able to delete my Yahoo UI text editor hacks.