Richard Das

On the complexities of the filesystem

I spent some time reading different comments and articles about the pro’s and con’s of the idea behind the iPhone OS, most specifically the fact that it is hiding the filesystem from the user. Personally I think it’s a great idea.

I began my computer experience way back, before the PCs, when you can sit in front of a green on black terminal and you could write COBOL (nasty!) on punched cards. Yes, I am old, but not as old as you might think… Anyway, UNIX was the king back then and you have to know how to interact with the filesystem since EVERYTHING is treated as a file. Open, read, write, close. Simple. However in order to access any information you first have to find it, or create it and save it and then find it. Eventually, after some time working with the OS you’ll find yourself with tons of files and while find (1) is a great tool to find files it is slow and heavy. You have to device a way to organize those files for future reference. It can get quite complex.

When the PCs appeared, DOS tried to emulate this (well, it tried to emulate this by copying CP/M but don’t get me started) and early on ran into the same problems. These problems were then moved into Windows. Actually most modern OS’s out there today share the same problem: The filesystem.

Yes, indexing technology and computing power got so good that using a program like Quicksilver or Mac OS’s own Spotlight makes your life easier, to the point that you don’t need to organize your files on subdirectories anymore. You can just find them at will based on name or content. That’s all good.

However…

You still need to create the files, name them, save them and remember with what application you want the files to be open with. Yes, most OS’s will remember (based on file extension, or content) the application you are supposed to use when opening those files. Still…

Now take the idea behind the iPhone OS: You open an app, you work with whatever data that app handles and you close the app. The next time you open that app, the data is there and you interact with it once more. You can chose to export that data into a common format (like text) or email it, or sync it, however you don’t have to remember where it is located, what name the file has or what app open that file. Simple.

We are starting to see more and more program for classic OS’s follow the same pattern. Programs such a the wonderful Notational Velocity (a simple and super cool text and note taking program) or my own Codex (a code and text snippet organizer). These programs don’t ask you to save the information, they do it for you, automatically. The next time you open the program, the info is there. You can start working right away.

Now, I know this might work out well for non technical or geeky people, but how about the professionals? The programmers, the sysadmins, the hackers?

The professional might need access to the filesystem and even need multitasking in order to debug or configure apps. That’s cool, you will have access to that, however it’s not going to be the default mode. It’s like a car, you don’t care about how the engine works, you just want to go from point A to B. If it breaks you can take it to a professional that has access to the engine (the filesystem) or become a hacker yourself and fix the engine.

Anyway, this subject has been in my mind for a couple of weeks already and I needed to make a brain dump here.

I hope you don’t mind.