In my last post i was commenting that the problem with computers today is in software.
Nevertheless, i realize this is not at all the best of times to be pointing to problems in software. With last week’s release of MacOS X’s last version (10.5 Leopard) and the beginning of the year’s launch of Windows NT 6.0 (Vista), it might seem that those software platforms are well established, complete and robust frameworks. Even the Gnu OS, strangely called Linux — to specify a major launch we could cite the recent Ubuntu 7.10 (Gutsy Gibbon) — has recently achieved a remarkable degree of consistency and usefulness.
And the question is, those platforms seem to be mature, reliable, and powerful. They seem to have solved all the problems.
I like the way a certain Paul Thurrott puts it: “OS X is so solid, so secure, and so functionally excellent that it must be getting difficult figuring out how to massage another $129 out even the most ardent fans.” He is much more kind to Vista, i should say, but overall he is also pointing out that it is mostly more of the same: “Microsoft’s job, now, is to convince you that its previous Windows system, XP, is not good enough. I don’t envy them that task.”
But the amazing, incomprehensible thing is that this is happening exactly while we are living the revolution of the digital, when finally computers are becoming the place for music and videos and your friends and your world-domination-plans and whatnot. This is the moment when computers are being put to deal with the juiciest parts of our lives, the ones we cherish and try to look cool about. This should be the time when computers would be learning new tricks.
Instead, they are just repeating the old ones.
Or learning to do those over the web.
One of my pet peeves is the search-everywhere explosion. Google Desktop and lookalikes, then MacOS Spotlight, then whatever it is called on Vista but it is exactly the same thing, and all the while a handful of engines and systems competing over different programming approaches in Gnu to do exactly the same thing.
Curiously, all of those things are attempts at doing exactly what BeOS did more than 10 years ago (and in a most sophisticated way)! The Be File System had all those things in the very filesystem — so that, for example, if the fs itself had been ported to Windows and to Mac and to Gnu you could share exactly the same search capabilities over all of your systems.
I do not think it is a bad thing that Operating Systems have become more stabilized, that they are not adding spurious and idiotic features at so fast a rate that whatever use your computer might had have will be sunken into loading times, but i do see a frightening adherence to stale ideas.
A small OS from a rogue company ten years ago could re-think how computers should work, and implement their new ideas into a system so good that it is being copied as of this day, and so why can’t the big players do the same?
It might at first seem that there is a major, big, fat resistance to change at work here. People got used to windows and pointers and icons and they want they where they were. That is, new ideas cannot substitute the old OS we have, they can only be added to them — and therefore be at best a complement and never a re-thinking in the level i am demanding here. But this is foolish. An addition can be radical — take Google, for example.
So that i do not sound as a complete BeOS zealot, let me state that there are plenty of similar examples in the recent past. OS/2, for one. Or the Palm. And on and on.
The crucial thing is not this or that technical decision, but the disposition to rethink what computers should do. Even to say, sometimes, that they should do what they are doing now.
My feeling is that the basic ideas that originated our present WIMP-with-some-twists use model of computers, have grown as far as they could go. We need a new model.
And the same arguments can be laid out about software outside the OS world. Take Adobe’s product line. There are almost nothing in their new releases that shouldn’t be reasonably hackeable out of their code base of ten years ago. The anecdotal evidence goes that, when asked for one thing that was really indispensable in the new Photoshop that didn’t exist in the 7.0 version, many will point the possibility to drag the image so that it’s border is inside the window border, to see the pixel in the corner. This, let’s face the facts, is neither something that needs great computational power nor complex algorithms.
What exactly do i propose to solve this situation? For now, nothing. My ideas are still flowing around.
“But for Christ sake, he better start looking!”