three computer things
Wednesday, May 5 2010
Some weeks ago a colleague introduced me to DropBox, an internet service that works through a background process on your computer. DropBox monitors certain directories on your local file system and if you put new things in them, it automatically copies their contents to some server on the internet. From there you can either access the files from the Web or share your directories with friends. It seemed like a useful service, though, aside from the integration with the local file system, it didn't seem like a big improvement over conventional FTP. Today, though, I made a miraculous discovery about DropBox when I went to upload a piece of pirated software (Acronis True Image) that I'd originally downloadead over a year ago using Bittorrent. Amazingly, though the installation was nearly 100 Megabytes in size, the "upload" only took about twenty seconds. At that point I realized something a little wonderful but also a little creepy: the "upload" was simply a "credit." The software on my computer had merely performed a checksum on the file I'd intended to upload, and then looked in the accounts of all the other DropBox users and found a file matching that checksum and then said that this file was available for download to my friends from me. Essentially, then, DropBox is a hyper-efficient cloud storage system, one where identical files in disparate accounts are stored only once. This makes for an ideal network for exchanging MP3s and wares, where most files have been uploaded by someone before. It's less useful for the sharing of original material (such as a collection of wedding photographs), because in that case time as well as your upload bandwidth are still tied up by an actual upload.
Still, DropBox is pretty useful. Those wanting to install in on their computers are encouraged to follow the links from this page, which, through a crude affiliate program, credits me 250 Megabytes of additional storage with every installation. Hey, I've been public radio without pledge breaks for you for the last 14 years; it's the least you can do.
Today while using Filezilla (the FTP client I have been using for many years), I realized I was experiencing a couple of annoyances, so I decided to create an account at Filezilla code central to request a couple features (something that is at least possible with an open source project). I wanted an option in the settings to allow me to turn off the "local site" pane (because I just drag and drop to and from explorer windows). And I also wanted option in the settings to automatically convert backslashes (the kind found in local Windows paths) into forward slashes (the only kind of paths found on web servers). It wasn't long after posting these requests that some pimply-faced nerd at Filezila dismissed these requests out of hand, arguing (for example) that because backslashes are accepted filenames in Unix operating systems, the backslash might prove problematic. Yeah, Asperger McKenzie, that's why I wanted it as an option in settings, for the 99.99% of people who never encounter such files. I'd be tempted to add this functionality to Filezilla myself, but the source code of that bitch (which I downloaded) is 18.5 Megabytes, and I don't even know how to compile C++ source on a Windows machine.
My final computer-related bitch of the day relates to Firefox autocomplete. The algorithm driving that is really poorly written; for weeks now the go-to page of my database visualizer has been tf.php, but yet on all the sites I'm working on (all of which employ my visualizer), Firefox helpfully suggests tableform.php. It doesn't seem to be aware of the fact that I've never once visited a tableform.php address in the last two weeks without encountering a 404 error. It doesn't seem like it would be rocket science for Firefox to handle autocomplete better than this.
For linking purposes this article's URL is:feedback
previous | next