Multiple Desktops

I know there are various ways to have multiple desktops on Windows too, but dad-gummit, they just don’t work as well as X’s built-in multi-desktop goodness. I never have just one program open; there’s always a browser for looking things up on desktop 1, my IDE of the moment on desktop 2, maybe GIMP on desktop 3, and a few terminal windows for convenience on desktop 4. There’s always a place. Switching is a quick ctrl-alt-arrow away, and that’s a good thing.

Linux Feels Faster

I don’t care about benchmarks (which usually go to Linux anyhow), I care about my own user experience. And in this case, Linux consistantly feels a whole lot faster than any Windows installation I’ve ever used.

I don’t know the technical reasons. Maybe it has something to do with Linux’s process-based architecture, but running multiple programs on Linux doesn’t seem to impact performance nearly as dramatically on Windows. Furthermore, when programs are thinking in Windows, the entire system is impacted; when some piece of processing takes a long time on Linux, it’s ok, because the rest of my UI is functioning at 100%.

That’s what I mean when I say it feels faster. If some program is taking too long to load, that’s alright, because I can browse the interweb while it’s loading and not even notice. Whenever I start up a program on XP or Vista it seems like the entire system locks up while I’m waiting for it to load.

The Terminal

The idea of ever using Windows Explorer again is nauseating. Who would want to do that when I have such an incredibly fast, efficient, and powerful tool at my disposal for file manipulations. Wildcard expressions to copy, move, rename, and otherwise manipulate files? Regular expressions even? I am so there.

Sure, Windows has DOS, but it offers no where near the flexibility – not to mention the UI.

Just today I converted a 40-page 300dpi pdf file into 40 resized web-ready .jpg images using a single command:

mogrify -density 300 -format jpg -resample 1024 *.pdf


I don’t even know how I would do that in Windows. Command-line image manipulation is pretty sweet.

How about grep? Grep is frigging fantastic. Compared with Windows search? Please.

The Community

You guys are swell, honestly. I have rarely had a problem that couldn’t be solved on the first page of Google’s search results. One of the nice things about open-source is that when it’s popular, it’s really popular, and people are eager to help.

Linux is great, and if you consider yourself a computer person you owe it to yourself to give it a try.

It is and always has been the popular thing to publish, um, very opinionated articles in blogs. I’m going to leave this up, but since this is the most-found page on this now-defunct blog (Although I still post from time to time at http://posterous.adambard.com/), I should say this: What follows is very bad advice.

If you ever need a hook to grab a link, just use a css class for god’s sake; jQuery et. al. make this ridiculously easy, and it was never that hard to begin with. Rel has a real meaning, if a generally-unused one, and there’s no sense using it when you really mean to use a class.

Carry on then. The rest of this post is the original in its entirity.

In HTML, <a> and <link> elements have a unique attribute, “rel.” Rel has practically no use in plain HTML these days, but with some creativity it can be an excellent hook for javascript events. This will provide you an unobtrusive way to change the behavior of these elements, without requiring “javascript:” links or onclick calls be written directly into the html.

Implementation:

var getElemsByRel = function(rel){
    aElements = document.getElementsByTagName("a");
    relElems = [];
    for (i = 0; i < aElements.length; i++) {
        if (aElements[i].hasAttribute("rel") && aElements[i].rel == rel) {
            relElems.append(aElements[i]);
        }
    }
	return relElems;
}

Case Study

Lightbox.js uses this to great effect: Any links with rel=”lightbox” activate the script, causing the image display to pop up with the target of the link as the image to display.

To accomplish this is very simple; for each element with the target rel attribute, one adds an onclick attribute that calls the function of one’s choosing, and – importantly – returns false, so that the link is not followed. As an added bonus, this technique fails extremely gracefully; if the javascript isn’t executed, the link is followed and the image is simply displayed in the browser.

Using rel in this case solves the problems of storing some relevant information in the link, forgoing the insertion of javascript into the markup, and providing some sort of semantic information in the markup as well.