I think this handles what we need, and can be implemented on current OSs.
In iTunes you don't manage the files. You manage metadata about the files. You can group songs arbitrarily (playlists), and each song can be in many groups. There's one pane where you see all the songs, so you can sort and find them without having to remember where you put them.
Now imagine it isn't just files, but each email message you get can have associated metadata and be placed into groups along with regular data files.
Then all you need is better search than iTunes, and maybe a more scalable way of dealing with the groups.
This is far better than Gelernter's approach because it gives you the ability to sort everything by time, but also to group in meaningful ways that aren't exclusive.
The other thing you need is that when you select a file, all the options for applications to open it should be easy. 80s tech associated a file with one application, but we're a bit more advanced and primitive than that. I might want to open an html file in one of several editors or word processors, and one of several browsers. I could want to open a Perl script in several editors, view the text in a browser, or execute the script.
Interestingly, you could make this compatible with current operating systems by creating directories for the groupings and putting links to the actual files in the directories. The way to hack namespaces is to put each actual file in a directory named for the id of the file. That way, for example, when you create or open a file in Photoshop, you get to name it something meaningful, but the gui replaces the file with a link, and puts the original in a uniquely-named directory somewhere behind the scenes.
A mindmap is really just a hierarchical outline -- in structure. It beats a regular outline for brainstorming because you can keep adding to each node. So part of the advantage is visibility and accessibility. You could use Word's outliner, but you would wind up scrolling, and unable to view the whole thing at once. Most mindmaps don't even go very deep. I hate the OS X magnifying dock, but it would be useful to have many lists on the screen, and have the one you point at magnify.
For a search engine at a single site, this is very useful. You watch the queries and results. If a page doesn't show up, but it should, you add the search terms to the keywords. If it shows up, but you don't want it to, what do you do? Create an anti-keyword field.
This is why XHTML requires all attribute values to be quoted.
Re:How Much Bandwidth Stylesheets Can Save You...
on
Slashdot Updates
·
· Score: 1
Excellent points. We need for someone to show the world how to do a website well. You would think Slashdot is a likely candidate.
The first thing is that ads should really add value. The current value-subtracting ads are completely ridiculous.
Beyond that, yes, cut down on bandwidth costs. mod_gzip is nice, but with a bit of cleverness you wouldn't have to gzip on the fly for most of the pages you send out. For example, put ads and what-not into iframes, so static elements are stored on the server gzipped, and dynamic elements are done separately. You might want separate versions of each page for NS4 and small or very old devices. Preprocessing is very cheap.
Even the length of urls matters when you have a lot of them on the page. Use the base tag, and then abbreviate like hell: href="/c.pl?s=99999&t=1&cs=0&m=t&p=9999999#9999999 " saves 39 characters per link, for example.
end of hijacking
on
More WTC News
·
· Score: 2, Insightful
The way I read the story of UAL Flight 93, passengers were about to regain control of the plane, and the last terrorist nosedived into a hill. A few people with small knives can't expect to control a planeload of passengers, if the passengers don't think cooperation will lead to survival.
Slashdot Mirror
I think this handles what we need, and can be implemented on current OSs.
In iTunes you don't manage the files. You manage metadata about the files. You can group songs arbitrarily (playlists), and each song can be in many groups. There's one pane where you see all the songs, so you can sort and find them without having to remember where you put them.
Now imagine it isn't just files, but each email message you get can have associated metadata and be placed into groups along with regular data files.
Then all you need is better search than iTunes, and maybe a more scalable way of dealing with the groups.
This is far better than Gelernter's approach because it gives you the ability to sort everything by time, but also to group in meaningful ways that aren't exclusive.
The other thing you need is that when you select a file, all the options for applications to open it should be easy. 80s tech associated a file with one application, but we're a bit more advanced and primitive than that. I might want to open an html file in one of several editors or word processors, and one of several browsers. I could want to open a Perl script in several editors, view the text in a browser, or execute the script.
Interestingly, you could make this compatible with current operating systems by creating directories for the groupings and putting links to the actual files in the directories. The way to hack namespaces is to put each actual file in a directory named for the id of the file. That way, for example, when you create or open a file in Photoshop, you get to name it something meaningful, but the gui replaces the file with a link, and puts the original in a uniquely-named directory somewhere behind the scenes.
A mindmap is really just a hierarchical outline -- in structure. It beats a regular outline for brainstorming because you can keep adding to each node. So part of the advantage is visibility and accessibility. You could use Word's outliner, but you would wind up scrolling, and unable to view the whole thing at once. Most mindmaps don't even go very deep. I hate the OS X magnifying dock, but it would be useful to have many lists on the screen, and have the one you point at magnify.
For a search engine at a single site, this is very useful. You watch the queries and results. If a page doesn't show up, but it should, you add the search terms to the keywords. If it shows up, but you don't want it to, what do you do? Create an anti-keyword field.
This is why XHTML requires all attribute values to be quoted.
Excellent points. We need for someone to show the world how to do a website well. You would think Slashdot is a likely candidate.
The first thing is that ads should really add value. The current value-subtracting ads are completely ridiculous.
Beyond that, yes, cut down on bandwidth costs. mod_gzip is nice, but with a bit of cleverness you wouldn't have to gzip on the fly for most of the pages you send out. For example, put ads and what-not into iframes, so static elements are stored on the server gzipped, and dynamic elements are done separately. You might want separate versions of each page for NS4 and small or very old devices. Preprocessing is very cheap.
Even the length of urls matters when you have a lot of them on the page. Use the base tag, and then abbreviate like hell: href="/c.pl?s=99999&t=1&cs=0&m=t&p=9999999#9999999 " saves 39 characters per link, for example.
The way I read the story of UAL Flight 93, passengers were about to regain control of the plane, and the last terrorist nosedived into a hill. A few people with small knives can't expect to control a planeload of passengers, if the passengers don't think cooperation will lead to survival.
Terrorism's first goal is to provoke indiscriminate retaliation.