::scr Ramblings of a Classic Refugee or How I Learned To Stop Worrying and Love OS X

[Matt Webb]

On Tuesday, February 5, 2002, at 02:40 PM, Simon Wistow wrote:

> On Sun, Feb 03, 2002 at 04:54:11PM +0000, Matt Webb said:
>> On this topic,
>> http://www.namesys.com/whitepaper.html
>> looks interesting (although I admit I've not been able to read it to 
>> the
>> end). Kind of xpath meets filesystems.
>
> Yes. Umm, that's why I quoted it as well :)

Darn, my lack of >3 second memory lets me down again.

>> But god yes, the current filesystem sucks. You only have to look at the
>> way email clients, mp3 programs, photo archives etc etc reimplement
>> their own filing systems.
> [ snippage ]
>> Yeah! That's views, or dynamic searches, or whatever you want to call
>> them. On that,
>
> They would rock. The big problems here are
>
> 1. performance -
> FS access would be *slow*. But memory is cheap and I've got clock cycles
> to burn so we'll let Mr Moore take care of that for us.

I don't know much about BeOS, but I think I've read something that 
claimed they'd fixed this problem? Maybe. I can't remember.

> Effectively DB file systems and also individual files would have the
> equivalent of DTDs (the file that describes an XML file for the buzzword
> protected) but for lovely shiny binary files. Al Snell would be proud.
>
> Essentially it would describe the data contained within the
> file/filesystem. When you transferred a directory over if you didn't
> already have the correct DTD it would get transferred over aswell.

It sounds gorgeous, but this is where my lack of computer history is 
really a problem.

A binary file with a separate definition document gets rid of the 
primary advantage of XML, namely that it can be edited in a standard 
text editor. In that case, why is a binary file like that better than 
files being objects? If they had specified accessor methods, they could 
pretend to be any file type they needed too.

>
>> So number one, get rid of all menu options in computer interfaces. And
>> dialogue boxes. And anything else that appears. All on-screen
>> manipulation should be done:
>> - by pulling a tool over to act-at-a-place; or
>> - by looking "closer" at the item to alter its properties
>
> This sounds suspiciously close to 'my' idea of document centric
> interfaces which I keep bringing up.
>
> have a look at http://www.thegestalt.org/scr/old/msg00023.html

Woo! We're riding the crest of the memewave!

Actually, I'm thinking about a different level.. 
Applications/documents/Opendoc etc, yes, brilliant. What I mean with 
tools and looking isn't at a scripting level, it's *physically* [can I 
say that?] moving the cursor to the same place as the thing if you want 
to alter property-of-thing. It's a look n feel thing.

>> [insert Whole Other Rant here.]
>
> Please do :)

Oh, I need to stumble onto the topic somehow. But briefly:

I will assert that there is some pattern[s] to the way the universe 
works. The closer an interface is to these patterns, the more 
familiar/intuitive it will feel. Therefore better.

The problem is twofold:
- the patterns are quite deep, so we've never really considered building 
them in to computers. I'm talking about patterns like cause and effect, 
weight, small changes only rarely having big effects, don't eat poison 
berries, that kind of stuff.
- with things like cars, they operate within the range of 
physical-reality so we get patterns for free. Stuff like turning a wheel 
having a proportionate effect on the turning circle (although I know 
that wasn't an inevitable innovation, and to alter the mirror position 
you *touch the mirror itself* (genius, you wouldn't get that in 
Windows). However with computers we're creating a whole other reality, 
and we're losing a whole load of things we never knew we had: peripheral 
vision, the idea of stretching to grab something, alter-at-a-point.

So computers have to be consciously designed to reintroduce these 
things, which is going to be hard. But it has to be done carefully and 
consistently.

Looking is one of those things. Consider a signpost at a crossroads. Far 
away, it's quite small. As it gets more important [you're closer to the 
crossroads], it gets bigger! What's more, the longer you look at it, the 
more detailed information you see, information you can neglect next time 
you encounter it because it'll be the same (you can tell because if it 
was different, the signpost would look new). Wouldn't that be great with 
dialogue boxes?

The problem is that signposts exist in the physical domain so that get 
stuff like perspective, and distance, and aging for free. And we're 
going to have to emulate that.


Hey, you remember I said "But briefly"? I lied.

	-mw