Slashdot Mirror

the reason is not-invented-here syndrome or a strawman argument against Unicode's ideograph system.

Asians complain about having to share codepoints for characters (see Han Unification). A lot of people think the whining is mostly racist in nature; however, there are a few legitimate complaints in there, for instance many characters that Han Unification forced to share codepoints actually have different glyphs depending on which language it's written in, even if they all shared a similar source. Unicode's official stance is that the application should identify the language being used and select the correct glyph for that particular character from a font designed for that language (which basically makes it impossible to create a single definitive Unicode font). Explanation of both the glyph problem and the not-invented-here issues here. More on glyphs.

As for the "major encodings" currently in use, Shift-JIS was basically forced onto them by Microsoft who took their existing JIS encoding (based on ISO-2022) and broke it. There's about 1000 websites of people telling the world just what they thought of that, yet not-invented-here or not, the Shift-JIS system sucks. Consider the fact that it uses two separate non-contiguous sets of codepoints (compare graphs of the non-unicode encodings here.

It's easy to say "not-invented-here", it's harder to admit that the foreigners had no clue what they were doing and broke it :P

it could just be a vocal minority

Of course it's a vocal minority, the majority just want their computers to work.

The truth and nothing but the truth

Man, that kinda reminds me of this.

This http://lfw.org/jminc/blog/http://news.bbc.co.uk/1/ hi/uk_politics/4284710.stm BBC news story now looks a lot more fittting.

Ohh, this is so much fun! http://lfw.org/jminc/suck/http://www.microsoft.com

http://www.microsoft.com/>Ohh, this is so much fun!

Then you would probally hate this.

This is one of my pet peeves.

Then you would probally hate this.

This is one of my pet peeves.

Blog blog blog blog. Blog blog.

Really cool, I remember using Shodouka - a proxy server that used the same technique to read Japanese webpages back in 1995 (long before browsers supported foreign languages). It was certainly a huge innovation back then, though with processor power as it is today, generating images on the fly like that might seem more trivial.

Here's his new CV.

Quite impressive. Apart from Xerox PARC, he's worked at ILM, Opera & Alias (Maya, Wavefront etc).

User Interaction Design for Secure Systems

by Ka-Ping Yee
Computer Science Department
University of California, Berkeley

Abstract:
The security of any computer system that is configured and operated by human beings critically depends on the information conveyed by the user interface, the decisions of the computer users, and the interpretation of their actions. We establish some starting points for reasoning about security from a user-centred point of view, by modelling a system in terms of actors and actions and introducing the concept of the subjective actor-ability state. We identify ten key principles for user interaction design in secure systems and give case studies to illustrate and justify each principle, describing real-world problems and possible solutions. We anticipate that this work will help guide the design and evaluation of secure systems.

Even better, let your network hand out IP's but take every user request and Jesusify or Malkovich them before sending them to the user.

Security through obsfucation.

grip

And in the real word, what babelfish does would be illegal. It is illegal for you to take the latest King novel and translate it into Tagalog and publish it

But babelfish does not take pages, translate them, and publish them. Babelfish is a program which mechanically translates text you give it. You view the output through a web interface, but that web interface is not "publishing", it is simply the interface you use to communicate with a network-distributed application. You definitely do have the right to take the stephen king novel and translate it to Tagalog in your own home, then read it. Babelfish can merely be thought of as a really fancy pen that somehow takes the difficulty out of that. That being said, what about those OCR-scanning "highlighter" pens they sell at computer stores that you roll over text and it translates it to other languages? That is different from babelfish how?

The law is a lot like programming. The users are all potential bugs, so you code around them.

Except those aren't bugs, they're people. And the law actually has a mandate to be fair.

That's easy. An illegal program changes the content or meaning of the page. ... The meaning of the page, derived from the meaning of the sum of its parts including the ad, is different.

What is "changing content"? By bugs in babelfish's translation mechanism, some small amount of content changing will inevitably occur whenever the mechanism makes a mistake. How do you legally define that? Intent of the program's author in changing said content? I guess, and your really could do that, but Intent is very difficult to prove sometimes.

That being said, what about the Malkovich Mediator? It unquestionably changes both the form and the content of the page it "translates". But look at how harmless it is, and how simple it would be to implement. Would you really call that illegal or unethical? I have seen Malovich Mediator programs that filtered text in a wierd (but upfront and clear to the reader) way in order to prove a political point, for example the drug slang thingy on Brunching Shuttlecocks (i cannot find the link). How would you reconcile banning a program such as that with free speech and expression concerns?

And as i said, i view programs as tools. What right have either you or the government to say what tools i may legally use to perform legal actions such as rewriting a copywrited text for my own use in my own home?

Oops, I forgot to change the link. Try it with microsoft.com

Try this with microsoft.com and read about CowboyNeal.NET

http://web.lfw.org/jminc/CowboyNeal/http://slashdo t.org/

This problem was solved five years ago by MINSE (a Medium-Independent Notation for Structured Expressions). It's not just a design, it's a working implementation. See http://www.lfw.org/math for details, or visit a demo page to see output (compare it to the HTML source of the same page).

Summary: enter equations like sane people do, e.g. "a*x^2 + b*x + c = 0" or "x = (_b +/- 'root(b^2 - 4*a*c))/(2*a)". Type them directly into your HTML; no need to run a converter to generate your pages. They appear in your web pages, look much, much better than LaTeX2HTML because they're antialiased, and anyone with a browser can see them without installing any software. Even text browsers work -- they get an ASCII art rendering!

Presented to the W3C but sadly ignored. At first it was rejected because they thought extensibility was unnecessary; then after they realized extensibility was critical, threw out their design and started over, they ignored MINSE because it wasn't XML. But there's one little point they missed: it actually works. Five years later, MathML is still vapourware -- and even if it did work, it would be completely unusable by teachers.

This problem was solved five years ago by MINSE (a Medium-Independent Notation for Structured Expressions). It's not just a design, it's a working implementation. See http://www.lfw.org/math for details, or visit a demo page to see output (compare it to the HTML source of the same page).

Summary: enter equations like sane people do, e.g. "a*x^2 + b*x + c = 0" or "x = (_b +/- 'root(b^2 - 4*a*c))/(2*a)". Type them directly into your HTML; no need to run a converter to generate your pages. They appear in your web pages, look much, much better than LaTeX2HTML because they're antialiased, and anyone with a browser can see them without installing any software. Even text browsers work -- they get an ASCII art rendering!

Presented to the W3C but sadly ignored. At first it was rejected because they thought extensibility was unnecessary; then after they realized extensibility was critical, threw out their design and started over, they ignored MINSE because it wasn't XML. But there's one little point they missed: it actually works. Five years later, MathML is still vapourware -- and even if it did work, it would be completely unusable by teachers.

1. Search for Linux
   
2. Click on Show matches (Cache) for ww.linux.com
   
3. Look at the Linux.com logo that comes up on that page. Look at the source, or even just go to "View Image" in the image's context menu (in Netscape).
   
   The URL is: http://198.186.203.57/i mg/template/linux.com-dropscreen.gif

The 'CPU power' chart disgusted me. If it really reflects the data in Kurzweil's book, I fear for humanity, because it was ludicrously rigged, and should have caused an outcry for such a highly regarded bestseller, yet apparently even the geeks didn't notice.

Yup, you heard me. The chart is rigged. Both with carefully selected points and with outright fabricated data.

First, point selection: Where are the mini's and mainframes and supercomputers of the 80s/90s? ("They aren't the most cost effective!" you cry) Okay, then where are the wang and other desktop (sometimes desk-sized) programmable calculators of the 60's/70's? The HP and other pocket calculators of the 70's/80's? ("They aren't the same thing," you argue, weakly.)

Second, bad data: "Fine," I say, "then do you really expect me to believe that a Hollerith tabulator took 3-30 hrs per operation? 10^(-4) to 10^(-5)ops/sec (according to the chart) = 10,000-100,000 seconds/op. In fact, there isn't a single computer capable of 1 op/sec until 1950 in the chart. Am I to believe that business spent millions on computers that were far slower than a moderately bright child using an abacus?

And how about the Apple II, the first personal computer I owned -- a 6502 used two clocks (out of phase) at roughly 1 MHz to run 500K single cycle ops/sec. I don't recall any common op codes that took more than 5-6 cycles, but there may have been one or two oddballs at 7-8 under worst case scenarios. (Most took 2-3.) Yet the Apple is listed as being roughly 30K ops/sec? (log(ops/sec) = 4.5)?

"Oh but we're talking about 64-bit adds" you argue, whipping out your abacus "...er, make that 'multiply's!"

I could invent justifications for each point on the chart, but by the time you're done, you'll realize, as I did, that the chart isn't worth the paper it's written on, and that it is utterly shocking that this hasn't been pointed out and ridiculted a million times by now.

3-30 hours for a single operation? C'mon!

Yeah. You can hardly miss that mile-long page title up there in the titlebar (much less if you bookmark it). "Foresight Institute Slash Server: Slashdot Like Automated Storytelling Homepage" ??? Sheesh. I wish they'd edit their HTML head tags down a bit. Other than that very minor caveat a cool site, though.
Check out the following URL for a nifty log-scale chart of CPU power plotted from Ray Kurzweil's book: here (needs Adobe Acrobat reader)
According to that chart, by around 2040 we'll be able to buy a human brain's worth of CPU for around $1000 (1999 $ - probably over a million bucks if current inflation holds ;-P) Assuming Moore's Law doesn't sputter out before 2020 like the pundits are predicting now. (well at least we'll have a mouse brain's worth by then - Pinky & the Brain gets a whole new lease on life).


#include "disclaim.h"
"All the best people in life seem to like LINUX." - Steve Wozniak

Congratulations to the Dreadling people for a nice piece of work.

I noticed some comments here about the speed and interface; you may want to check out Escape, a demo i wrote in 1997. I believe it was the first real-time texture mapping program ever created for the Pilot.

Get it from http://www.lfw.org/pilot/.

It has about twice the frame rate of Dreadling; its graphics are simpler (no monsters, no greyscale) but i think the maze looks nicer, and the faster response is a big win.

Steer by dragging the pen around the screen, rather than using the buttons.

-- ?!ng

What you're talking about is called a mediator .

I've been doing this for years. In 1995 i came out with Shodouka, which is a mediator that replaces JIS codes with GIF images so anybody can see Japanese text even if they don't have fonts installed. In 1996 i created MINSE, a semantic expression language with a mediator that lets anybody see math expressions directly embedded in web pages. MINSE was special because, like this so-called "transcoder", it would translate the equation into an appropriate form for the browser: use Netscape, and you get nice antialiased GIF images of the equations; use Lynx, and you get a good attempt at ASCII art. It is still the only way to easily put math on the Web that anyone can view.

In 1997 i did Crit, which enabled anybody to make public annotations on any Web page for the first time. You might want to check that out too (source code is available). It makes all links bidirectional and allows you to make links from your document to a specific phrase in the target document. As a side benefit it shows you some useful metadata about the document which browsers often hide. Again, since it's a mediator, anyone can view or create annotations using any browser -- people running Lynx can attach annotations to anything too.

The whole idea of "Web middleware" has been around for a long time. I'm pretty sure i was first, but many others have done similar things since then (e.g. Anonymizer, babelfish, and so on). Rohit Khare wrote a nice paper in 1998 summarizing the idea and various applications.

-- ?!ng

What you're talking about is called a mediator .

I've been doing this for years. In 1995 i came out with Shodouka, which is a mediator that replaces JIS codes with GIF images so anybody can see Japanese text even if they don't have fonts installed. In 1996 i created MINSE, a semantic expression language with a mediator that lets anybody see math expressions directly embedded in web pages. MINSE was special because, like this so-called "transcoder", it would translate the equation into an appropriate form for the browser: use Netscape, and you get nice antialiased GIF images of the equations; use Lynx, and you get a good attempt at ASCII art. It is still the only way to easily put math on the Web that anyone can view.

In 1997 i did Crit, which enabled anybody to make public annotations on any Web page for the first time. You might want to check that out too (source code is available). It makes all links bidirectional and allows you to make links from your document to a specific phrase in the target document. As a side benefit it shows you some useful metadata about the document which browsers often hide. Again, since it's a mediator, anyone can view or create annotations using any browser -- people running Lynx can attach annotations to anything too.

The whole idea of "Web middleware" has been around for a long time. I'm pretty sure i was first, but many others have done similar things since then (e.g. Anonymizer, babelfish, and so on). Rohit Khare wrote a nice paper in 1998 summarizing the idea and various applications.

-- ?!ng