bangersandmash

That’s not perfect, but it’s clearly better than the actual tab bar design in the Safari 4 public beta. Consider: with the previous tab design, if you wanted to move a window you dragged the window, and if you wanted to move a tab, you dragged the tab. Now in Safari 4, if you want to move the window you drag a tab, and if you want to move a tab you drag the small grippy strip at the far right edge of a tab. This is more abstract, indirect, and worse. Chrome’s tab design suffers none of these problems. from: Daring Fireball

I suspect that the “problem being solved” is that there was nothing about the old tabs that indicated you could move (or do anything with) them. Though the old title bar is gone, the behaviour is the same as before — click anywhere on the top of the window (except the grippy bit) and the window moves. Click the grippy bit and voilà, suddenly the masses know they can move their tabs about at will.

Here’s my time vampire of the day:

Flashbake

I’ve been using SVN until now, but git has intrigued. Flashbake’s ability to track context as well as product seems really interesting. More to follow.

Update: 2009-03-05 Fail. Just couldn’t get it to work. Oh well. I blame my lack of terminal-fu. Interesting project, though.

Spent last night reading a few articles on the relationship between searching and serendipity. This morning’s indexed seems to fit the bill perfectly

The basic tools that I use every day are

1. A browser (leaning towards Opera at the moment…)
2. Skim for reading PDFs
3. Bibdesk for citation information
4. Devonthink 2.0 for collecting everything (notes, images, etc) into a single place and making them searchable in a wiki-like environment
5. Textmate

Basically, I read material using the browser or skim. If something — an entire document or just a passage — it gets copied into Devonthink for later. If it is really useful its citation information gets added to my bibdesk bibliography. All writing is done in Textmate because it integrates so well with Bibdesk. Because various bookmarklets, OSX services and other trickery allow me move data between the applications with a click of a mouse or a keyboard shortcut it feels as though these five tools behave as a single application. Perhaps one day I’ll get around to compiling a video of what this looks like in practice.

Click image to view larger original version.

Having made many-a-flow chart in my day I just had to share this.

Storage space is cheap these days. Some are even giving it away for free. For me, however, file size is an important consideration. I use a Macbook as my primary workhorse and I want as much information as possible to be at my fingertips at all times. This means that I don’t want to be working with external drives, CD backups and ‘cloud storage’ (except for purposes of backup, of course) I’ve got an 80 gig drive so I should have lots of room for my materials but once you install a few programs, take a few hundred pictures and put a few gigs worth of MP3s onto the disc you’re left with a lot less room to play with. This is why I’ve gotten fussy about what I save from my research.

Web pages are particularly nasty wasters of space. I recently saved a single article from the New York Times website that took up over 1100k! That’s more than a megabyte for a single article. After I’d taken 20 seconds to cut out the adds, javascript and other useless, irrelevant material that was saved in the web archive I was left with a file of 7k. The New York Times stuck me with 1093k of unnecessary material. Now multiply this across thousands of files and realize just how much drive space could have been wasted.

That’s why I’m thankful for Aardvark and make.text.

Aardvark is a bookmarklet (or plugin if you use firefox) that allows you to highlight the part of the page you’re actually interested in an automatically (or sometimes manually) get rid of unwanted content. This is usually more than enough to get rid of much of the crap that comes along with most web pages. However, make.txt takes this a step further by converting the remaining text on the page into markdown. This step is pure nerdery but it does an excellent job of preserving

• the link to the original source,
• all of the remaining links on the page
• the semantic structure of the original document.

What I’m left with is a small, easily searchable file of unformatted text that can easily be imported into other applications for editing and quoting. And no ads.