> they have released all their search algorithms to the scientific community
As far as I know only the original version of the google algorithm was released when Google was still an academic project. However, though that algorithm is still the basic principle, all modifications to prevent abuse through link farms, duplicate contents etc. are not public.
Actually, terahertz imaging is much more advanced than what is shown on the pictures, at least if you use a terahertz emittor instead of depending on the minute amount of thermally emitted radiation. Take a look at the Teravision project or the pictures at the terahertz research group at the Rensselaer polytechnic institute.
With passive terahertz imaging, you have to face the problem that basically everything, including your detector, is emitting terahertz radiation. With a well-designed pulsed terahertz source, as in the links, you are much less sensitive to thermal noise. Unfortunately, it requires quite a bit more equipment, most importantly a femtosecond pulsed laser system which is too expensive (at least USD 150,000), bulky, and fragile for use outside a laser laboratory.
Microsoft: 'Windows' in software name is not allowed.
Friday, November 29, 2002 - Microsoft has asked manufacturers of
software with the name 'Windows' to choose a different name.
by Maarten Reijnders
This appears after a quick tour along the manufacturers of
software listed on the lindows.com site. The makers of 'Windows Spy'
and 'Windows Backup Wizard' received a letter from Microsoft's
lawyers, as they confirmed to WebWereld.
Earlier, the maker of 'Windows Commander' had changed the name of its
program into 'Total Commander'. He did so after having received a letter
from Microsoft's lawyers.
The Russian maker of 'Windows Backup Wizard' decided not to conform to
the request he received from Microsoft in July. Alexander Tchirkov of
Windows Backup Wizard: "I received a letter of Microsoft's lawyers
with the recommendation to change the name of my program into
'Backup Wizard for Windows(R)'."
"I am not planning to change the name of my program. But Christian
Ghisler (maker of Windows Commander, MR) had already been forced to
change the name of his software, so anything is possible", says
Tchirkov who points out that Windows isn't a registered trademark in
Russia.
Software manufacturer Sureshot, however, did decide to change the
name of the program 'Windows Spy' into 'Farsighter'. "Microsoft
appears to possess the term 'Windows'", Jon, of Sureshot, sighs.
Windows Commander, Windows Backup Wizard, and Windows Spy are
mentioned on a list composed by the manufacturers of the Linux-based
operating system Lindows. The company plans to use this list in a
juridical procedure that Microsoft has started against Lindows.
Microsoft believes that 'Lindows' is too similar to 'Windows' and
demands therefore that Lindows stops using that name. In March, the
judge ruled that Lindows is allowed to use the name until the final
decision of the court.
Since the list at Lindows.com was published, it appears that
not only the names of Windows Commander and Windows Spy have been
changed, but also the ones of Windows Network Booster and
Windows Personalizer 2000. At least, the programs are no longer
available at their original name at Download.com
>An auxilary problem is artificial intelligence. Its seem obvious that this can be done by us computer geeks. But 55 years of effort have had disappointing results.
The unability of creating artificial intelligence might be more because of a lack of processing power than because of a fundamental problem.
In an article in Scientific American, it was estimated that the
human brain has a processing capacity equivalent to around
10^12 instructions per second. We just need to wait until the computers are factor 1000 faster than today to get somewhere, which would take around 20 years according to Moore's law.
>The first thing anyone should learn with a computer is how to code using a very, very simple language. I used MS QuickBASIC.
Agreed. But how do you do that on Linux? Linux and friends are supposed to be well suited for programmers, but C and Perl are not really for kids (or beginners in programming anyway).
There are a few Basic interpreters for Linux, but it looks like all of them are of the old-fashioned "interactive linenumber-oriented interpreter thing". Not the userfriendly editor+compiler like in Qbasic (or Turbo Pascal 6). (A beginner would be scared by separate editors and compilers etc.) And then, where to find a good Basic tutorial that is adapted to the particular basic dialect?
Wouldn't it be more robust if the server superimposes a verification signal on top of the working units? Something like a sinus wave with a certain frequency and amplitude. It is a small effort for the server to add such a signal, but the client has to do the full signal processing to find it back in the data.
Returned working units which do not contain the known verification signal can then easily be discarded as cheats.
Of course, this would not help against cheaters resubmitting the same data over and over again, but at least it would be sure that the returned data is not forged.
Exactly. However, the right tool may depend on the person who does
the job. A scientific environment, that owes is existence to the
presense of people who want to try new ideas is no place for political
debates on operating systems. If a scientific employee can produce
more publications if does the data analysis and paper writing on a
system that he is comfortable with, why would you deny him that right?
> write your reports and papers in Word so that the coauthors
can read and modify your text
It might not at all be desirable to have co-authors modify the
digital text of a manuscript. The iterative process of revising a
manuscript becomes much more tedious if it is not clear where
co-authors changed the text (and possibly introduce errors in an
attempt to "improve" the readability).
So will this increase or decrease the push to upgrade to new versions of Office? People are more or less forced to upgrade to new Office versions partly because you can't read those Word documents with old versions of the software which people e-mail to each other for some reason (this viral behaviour is thus propagated as the upgraded user starts to mail his documents around).
Would Office11 not running on older computers (a) boost the sales of w2k as well, or will it (b) have the opposite effect of being unsellable?
>Maybe they can predict the weather a couple of days with this.
That will only address the problem of inaccurate models. It will not decrease the problem of sensitivity to noise in the input data (the butterfly effect), which fundamentally limits the prediction to a week or so. To reduce the noise problem, we need more sensors all over the earth and the oceans.
>He just said you shouldn't spend all your time looking for any patent for an idea that someone may of thought of independently
Does anyone have an idea how much effort it would take to find out whether an algorithm is already implemented by someone else? It's often quite hard to find an algorithm in a book (at least in many books) that is known to contain the algorithm if you don't know in which chapter to look.
Is sed Turing complete?
on
SedSokoban
·
· Score: 1
If sed provides the essential capabilities for programming such a game, it probably satisfies the criterion for being Turing complete. It can be proven that any Turing-complete device can in principle do what any other Turing machine or computer can do.
Hence, it is probably possible to write an processor emulator in sed which can run an operating system and any software that exists in that operating system. It would only be a bit slow.
Dvorak is actually not much faster than QWERTY (neither slower), but as an experienced Dvorak typist, I can assure that it puts LOTS less strain on your fingers.
However, LaTeX with all the backslashes and C/C++/Perl all remain cumbersome, because Dvorak was optimized for English language and not for computer programming languages that use almost every printable character in the ASCII set. I'm just in doubt whether dollar ($) and backquote (`) have any obscure meaning in C++
>And, yeah. XML is just the neatest, niftiest, coolest thing to hit the planet since Apple ][e BASIC... But, really, is stuff like this really needed?
No, Commodore-64 basic was better! No, ZX-Spectrum BASIC was the best! No, Tandy! No, IBM's GWBasic, because it had an ELSE statement! No, INTBASIC on the Apple ][!
Why would one want standards if it's perfectly clear which is the best?
>No, this is not as horrible as Telnet. A VERY VERY small percentage of people that might be sniffing your network have the skills, motivation and time necessary to analyze your keystroke timing and use that information to any useful extent.
I do not agree. Just one skilled person needs to write a script and some instructions, after which the script kiddies will can use it 'en masse'. This is what happens to most security holes; finding out where the buffer overflows are, and then trying to construct the machine code that will execute exactly the right instructions is quite hard. But once it is automated in a script, you don't need special skills.
Most comments here have a tendency towards "If you want to have fast
downloads, get a faster modem", implying that it a functional webpage
should include tons of graphics.
If one wants flashy blinking stuff, it's much more effective to
watch MTV. Internet is for exchanging information, and most images do
not enhance that purpose. Is it easyer to find information in a
tabloid-style newspaper with everything on the page screaming for
attentention in all colors of the rainbow, or in a serious newspaper?
Then why does everybody want to have a webpage look like a tabloid?
Both designers and visitors, apparently.
Images don't scale nicely with screen size or default font size,
you can't do a text search in an image, they take lots of time to
download unless you are willing to pay USD 100 per month for a private
broadband connection. On Useit.com, you can find a large
archive of tested information on usability of web sites that support
these statements.
The only real purpose that most images on web pages serve, is to
obscure the structure of a website, such that visitors will spend more
time to find out what it's all about --- and see more ads.
>Of course I wish more of the media coverage would
criticize Microsoft for making holey software that
allows these worms to propagate so easily, but you can't always get what you want.
This is not specifically a Micro$oft problem. There have been plenty of security holes in Linux/Unix daemons (e.g. imap, pop) that allowed crackers to acquire root permissions. Imagine what would happen if someone wrote a script to test random hosts for vulnerability, then acquire root permissions on them, and subsequently install itself everywhere.
Maybe Linux users are too kind to do such a thing to their brothers;-)
Slashdot Mirror
As far as I know only the original version of the google algorithm was released when Google was still an academic project. However, though that algorithm is still the basic principle, all modifications to prevent abuse through link farms, duplicate contents etc. are not public.
With passive terahertz imaging, you have to face the problem that basically everything, including your detector, is emitting terahertz radiation. With a well-designed pulsed terahertz source, as in the links, you are much less sensitive to thermal noise. Unfortunately, it requires quite a bit more equipment, most importantly a femtosecond pulsed laser system which is too expensive (at least USD 150,000), bulky, and fragile for use outside a laser laboratory.
Friday, November 29, 2002 - Microsoft has asked manufacturers of software with the name 'Windows' to choose a different name.
by Maarten Reijnders
This appears after a quick tour along the manufacturers of software listed on the lindows.com site. The makers of 'Windows Spy' and 'Windows Backup Wizard' received a letter from Microsoft's lawyers, as they confirmed to WebWereld.
Earlier, the maker of 'Windows Commander' had changed the name of its program into 'Total Commander'. He did so after having received a letter from Microsoft's lawyers.
The Russian maker of 'Windows Backup Wizard' decided not to conform to the request he received from Microsoft in July. Alexander Tchirkov of Windows Backup Wizard: "I received a letter of Microsoft's lawyers with the recommendation to change the name of my program into 'Backup Wizard for Windows(R)'."
"I am not planning to change the name of my program. But Christian Ghisler (maker of Windows Commander, MR) had already been forced to change the name of his software, so anything is possible", says Tchirkov who points out that Windows isn't a registered trademark in Russia.
Software manufacturer Sureshot, however, did decide to change the name of the program 'Windows Spy' into 'Farsighter'. "Microsoft appears to possess the term 'Windows'", Jon, of Sureshot, sighs.
Windows Commander, Windows Backup Wizard, and Windows Spy are mentioned on a list composed by the manufacturers of the Linux-based operating system Lindows. The company plans to use this list in a juridical procedure that Microsoft has started against Lindows.
Microsoft believes that 'Lindows' is too similar to 'Windows' and demands therefore that Lindows stops using that name. In March, the judge ruled that Lindows is allowed to use the name until the final decision of the court.
Since the list at Lindows.com was published, it appears that not only the names of Windows Commander and Windows Spy have been changed, but also the ones of Windows Network Booster and Windows Personalizer 2000. At least, the programs are no longer available at their original name at Download.com
>10^12 instructions per second. Oops, I mean 10^14, or 100.000 times faster, 35 years of Moore's law.
The unability of creating artificial intelligence might be more because of a lack of processing power than because of a fundamental problem.
In an article in Scientific American, it was estimated that the human brain has a processing capacity equivalent to around 10^12 instructions per second. We just need to wait until the computers are factor 1000 faster than today to get somewhere, which would take around 20 years according to Moore's law.
Agreed. But how do you do that on Linux? Linux and friends are supposed to be well suited for programmers, but C and Perl are not really for kids (or beginners in programming anyway).
There are a few Basic interpreters for Linux, but it looks like all of them are of the old-fashioned "interactive linenumber-oriented interpreter thing". Not the userfriendly editor+compiler like in Qbasic (or Turbo Pascal 6). (A beginner would be scared by separate editors and compilers etc.) And then, where to find a good Basic tutorial that is adapted to the particular basic dialect?
Returned working units which do not contain the known verification signal can then easily be discarded as cheats.
Of course, this would not help against cheaters resubmitting the same data over and over again, but at least it would be sure that the returned data is not forged.
Exactly. However, the right tool may depend on the person who does the job. A scientific environment, that owes is existence to the presense of people who want to try new ideas is no place for political debates on operating systems. If a scientific employee can produce more publications if does the data analysis and paper writing on a system that he is comfortable with, why would you deny him that right?
> write your reports and papers in Word so that the coauthors can read and modify your text
It might not at all be desirable to have co-authors modify the digital text of a manuscript. The iterative process of revising a manuscript becomes much more tedious if it is not clear where co-authors changed the text (and possibly introduce errors in an attempt to "improve" the readability).
Would Office11 not running on older computers (a) boost the sales of w2k as well, or will it (b) have the opposite effect of being unsellable?
That will only address the problem of inaccurate models. It will not decrease the problem of sensitivity to noise in the input data (the butterfly effect), which fundamentally limits the prediction to a week or so. To reduce the noise problem, we need more sensors all over the earth and the oceans.
>He just said you shouldn't spend all your time looking for any patent for an idea that someone may of thought of independently
Does anyone have an idea how much effort it would take to find out whether an algorithm is already implemented by someone else? It's often quite hard to find an algorithm in a book (at least in many books) that is known to contain the algorithm if you don't know in which chapter to look.
If sed provides the essential capabilities for programming such a game, it probably satisfies the criterion for being Turing complete. It can be proven that any Turing-complete device can in principle do what any other Turing machine or computer can do.
Hence, it is probably possible to write an processor emulator in sed which can run an operating system and any software that exists in that operating system. It would only be a bit slow.
However, LaTeX with all the backslashes and C/C++/Perl all remain cumbersome, because Dvorak was optimized for English language and not for computer programming languages that use almost every printable character in the ASCII set. I'm just in doubt whether dollar ($) and backquote (`) have any obscure meaning in C++
No, Commodore-64 basic was better! No, ZX-Spectrum BASIC was the best! No, Tandy! No, IBM's GWBasic, because it had an ELSE statement! No, INTBASIC on the Apple ][!
Why would one want standards if it's perfectly clear which is the best?
I do not agree. Just one skilled person needs to write a script and some instructions, after which the script kiddies will can use it 'en masse'. This is what happens to most security holes; finding out where the buffer overflows are, and then trying to construct the machine code that will execute exactly the right instructions is quite hard. But once it is automated in a script, you don't need special skills.
downloads, get a faster modem", implying that it a functional webpage
should include tons of graphics.
If one wants flashy blinking stuff, it's much more effective to
watch MTV. Internet is for exchanging information, and most images do
not enhance that purpose. Is it easyer to find information in a
tabloid-style newspaper with everything on the page screaming for
attentention in all colors of the rainbow, or in a serious newspaper?
Then why does everybody want to have a webpage look like a tabloid?
Both designers and visitors, apparently.
Images don't scale nicely with screen size or default font size,
you can't do a text search in an image, they take lots of time to
download unless you are willing to pay USD 100 per month for a private
broadband connection. On Useit.com, you can find a large
archive of tested information on usability of web sites that support
these statements.
The only real purpose that most images on web pages serve, is to
obscure the structure of a website, such that visitors will spend more
time to find out what it's all about --- and see more ads.
This is not specifically a Micro$oft problem. There have been plenty of security holes in Linux/Unix daemons (e.g. imap, pop) that allowed crackers to acquire root permissions. Imagine what would happen if someone wrote a script to test random hosts for vulnerability, then acquire root permissions on them, and subsequently install itself everywhere.
Maybe Linux users are too kind to do such a thing to their brothers ;-)