Yahoo denies that its new initiative is a declaration of war on Google. Eric E. Schmidt, 47, Google's chief executive, also says the two companies are still allies. But relations are strained.
Is he being diplomatic, or is just in denial. I wasn't that impressed with Eric's performance at the helm of Novel (IMHO the mere fact that he took the job showed incompetence). And why did Serge Brin choose Eric Schmidt as CEO? 'cuz it was the only candidate that had been to burning man'.
As Marc Andressen said, once you are big enough, you appear in the radar scopes of the big guys, and your true mettle is tested. So far the comments above from Eric Schmidt and Serge Brin give me little confidence....
- an index of even just 10% of the web would satisfy most people, as long as it is the right 10% and is searched effectively.
Altavista tried this, Lycos tried this, excite tried it too, and guess what? users did mind. They would get really pissed that they couldn't find a web page they had seen earlier.
Traditionally, the Economist has been one of the earliest media outlets to get technology. The first printed press reference to the Internet outside the tech press is from the Economist, ditto for Linux, but this time around they are way off the mark.
The reason Linux is so popular is not that is free. BeOS is also free. Linux is successful due to convergence of many different factors:
(1) Free (2) Open source (3) Unix compatible (4) X-windows (X11) compatible (5) designed for x86 (yes it runs on many other chips, still Linux is an x86 project from the get go) (6) Multiple vendor supported (7) Plenty of third party support
Moreover each of these things feed of each other. That is why Linux is so popular.
Sonny, listen to me. We geeks have a tendency to either be incredibly cheap or get so overly excited about finally having a girlfriend that we go overboard on gifts.
Giving her an expensive gift will not improve your image in her eyes. To the contrary, it will send the message "I'm so desperate I'm willing to shell $3K to impresss you". Knowing she has you wrapped around her finger will encourage her to search for another potential boyfriend, as she knows there is a fail-safe option at home.
Don't get me wrong. Be nice to your GF but don't over do it, and occasionally you might want to hold back a little on purpose so she yearns for you.
Save the $3k and promise to visit her in a few months. If she really loves you, the thought of having you in Japan (for a _visit_, you dirty minded folks) would be gift enough for her.
Do you know the history of C? Do you know why it was created?
Of course. Algol and Pascal were introduced at the same time yet do not have this obnoxious behaviour. I bet you didn't know that.
You wouldn't want to accidently activate the whole set of floating-point division routines unless you really wanted to.
Let's assume for the sake of the argument that is true, why not fix it now? For example, sqrt(2) used to evaluate to garbage. The latest version of C does the proper casting to float and returns 1.41...
I really don't see why so many/.ers have come to assume that C/Unix/whatever are written in stone and cannot possibly be improved (contrast this with perl apocalipsis and python 3, which are efforts starting from the assumption that lots of things could be fixed).
People are so used to this design flaw that they can no longer tell that it is there. What would you say if I wrote a language in which the name of rebooting subroutine was say, "print"?
Well the code in my.sig is no different. It looks like standard schoolbook arithmetic, hence the outcome should be a resonable approximation of it.
Yes, Java has the same flaw as it inherited several C/C++ design flaws. In fact the bug can be traced back to Fortran.
Pascal and the new version of Python, algol, modula and others don't have this flaw as they use div for integer division.
Good, safe, secure and efficient code is generated by educated programmers who are aware of what they're doing.
Ah the typical real men need no bounds check argument. This is, of course, a bogus argument.
In real life, people cannot be expected to be extra careful day in and day out. It's just not humanly possible. The long history of buffer overrun exploits proves this.
Seriously, with the kind of computational power that is now available even in a bottom-of-the-line gray-box PC the time has come to retire mainstream languages that do not have array bounds checking built-in.
<rant> We are placing the entire computing infrastructure of the nation at risk in the name of a few measly clock cycles that would be spent performing bounds check.
Engineers know that the only way to have realiable systems is to have substantial safety margins and backup systems. This is a lesson that we software developers ignore at our own peril. </rant>
However, IBM's contributions to Linux pale, or become transparent to be more accurate, in comparison to the dedicated hackers from all over the world who contribute to Linux.
If IBM went to the trouble of hiring companies to improve Linux then some development that they wanted to see through wasn't being done. How important it was? important enough for IBM to spend millions of dollars on this.
Keep im mind that some of the work that makes an OS entreprise strength is nitty gritty optimization that the hackers tend to skirt, just like they avoid quality GUI issues (designing tons of star trek inspired skins==cool, designing *one* usable skin for the entire Linux desktop==uncool).
While I have no idea if SCO has a case or not, I see that many here assume IBM had nothing to do with making Linux enterprise stable, and scoff at SCO's claim.
Yet, if you take the time to google the web you'll find that IBM dedicated an entire internal group to Linux and hired several external companies during 1999-2001 with the sole purpose of making Linux entreprise strength (even Linus has said so).
Now, to be clear, this does not yet prove that any illegal transfer of technology took place (and I doubt SCO will be able to prove it, IMHO they are fishing hoping to find the smoking gun during discovery), but it does verify one of the main three claims from SCO.
The fact that you were able to put together this software package so swiftly leads me to conclude that you received assistance from IBM. This assistance infringes on my patent for a "device to increment accumulator by value unit value". My lawyers have filed papers suing IBM for
They have the resources to fix their security problems. They just haven't. Even when they know about a problem, it often goes unfixed for months, if not longer.
Sure but this is a different problem: Microsoft is unresponsive.
So let's not confuse quality of released code and responsiveness of manufacturer.
Personally, I agree with your comment on M$ unresponsiveness. We once found a bug in the optimization routine of Microsoft C. We had the platinum (or whatever the name is) support package with M$ which included 'face' time with "experts" so we sent an e-mail with the bug report.
Their answer: there won't be a patch forthcoming, it won't be fixed any time soon, it is not even clear if it will be fixed a year and half from now when we release the new version.
This is in contrast with other software manufacturers for which we would often get patches sent to us within a month or so of reporting a bug.
Ever hear anything good said about MFC? What about ATL? How about their OLE implementation?
Actually, Miguel de Icaza of gnome fame speaks highly of OLE... but that is besides my point.
I'd never argue that M$ puts out quality code. My point is that M$ code is crappy but no worse than that put out by competitors.
This is based on my use of several different platforms during the day (Unix, Linux, Apple, WinXP/2K) as well as software from many different manufacturers. Microsoft products do not stand out as being particularly more crash prone compared than the rest.
p.s. In fact currently the only piece of software that currently stands out as particularly buggy is my beta version of Mozilla 1.3b, which is quite understandable....
At the same time, from a security standpoint things are as bad as ever. Of all the machines on my network here, the only ones that have ever been compromised are the Windows boxes.
I agree that this is the case. Yet, a sensible explanation of why this is so is that more people write Windows exploits than Unix. Hence more flaws are found. The old adage of OSS is: debugging is fungible. Well so is cracking. Let's call that Alomex Law "cracking is fungible", as a consequence the most popular platform will be the most cracked. In fact as Linux continues to gain ground we'll see an increase in exploits (side note: about a year ago all our Red Hat boxes were cracked here).
In simple terms, if my goal as evil-cracker is to maximize disruption why would I spend hours pouring BeOS code that would give me access to four computers, when I can build upon readily available cracking utilities that can give me access to 90% of the world's computers?
Changing your conception of reality
on
Defining "Planet"
·
· Score: 1
It seems to me that one of the reasons this issue is so controversial is because it would mean we have to update our conception of reality, be it dropping Pluto or adding Ceres, and people just hate that.
I remember back in the 70s when telling people that Neptune was farther out (at that time) than Pluto, they would get very upset and push aside the SciAm diagram explaining the intersecting orbits. These are people who would otherwise be found a kid talking about the stars and trips to the moon quite endearing.
And ten years later at the next high school reunion, everyone will feel sorry for you since and your class mates (while poorer) had time to live their life (have family, friends, and kids) while you worked all your free time away at Microsoft.
Feel sorry for you?? Clearly you haven't been to your ten year school reunion.
They are the best ego trip ever for us geeks!
Turns out that the bimmer you are driving, which is non-distinct among the Benz kompressors and Porsche's carrera in your company parking lot is the best car of those in your class. It also turns out that you are making the most money. You are wearing the nicest suit and all the chicks that use to shun you drool over you. Odds are that you'll get some by the end of the night.
So if it's such a great place to work, and they're woo'ing all of the best intern material, WHY do they continue to produce such crap code?!?
Do they really or is it just a urban myth?
People used to complain about how buggy IE was, but when finally Netscape's code was released for Mozilla turned out it was no better. So much so that Mozilla had to discard Netscape's code.
Another example, in 1994 I was routinely using a Mac, and it would crash ever one or two hours (cooperative multitasking anyone?). Back then you could read anywhere how bad an unstable Win 3.1 was, but you'd never hear a peep from the Apple camp.
Right out of university? Without even letting them join the unemployed lines for a while? The _gall_
2) condition them to believe working extremely long hours and weekends is "normal".
As compared to all other programming jobs in Sillicon Valley, where they want you there only from 9 to 5, and they send you home if you are there past 5.............NOT!
3) condition them to believe that if you're a real professional then your work is more important than socialising with your friends and spending time with your family.
Huh? M$ and other software companies only hire nerds. They come preconditioned to believe that. You don't believe me?
4) pay them relatively low salaries, but promise big ones in the future.
In your dreams. M$ used to pay below average salaries but give generous stock options, for a way above average salary. A couple of years back when the stock price stopped moving up, M$ instituted a pay increase across the board...
5) give them free pop, sweeties and toys.
This is supposed to be bad, or you are just jealous?
6) See how far you can push the suckers!
While all other companies only have your best interests in mind right?
No, it's wrong. You merely have the blind leading the blind.
Bullshit. Implementing a fix that is not QA'd and fully tested should never be done. Clearly you have no idea how to develop robust, reliable software.
The real solution is to indicate whether a DLL is a bug-fix release or whether it represents a significant and incompatible change to the APIs.
This doesn't work, as often application developers detected the original bug, e.g./* string terminated by \0\n\0 due to DLL bug */
then the fix gets released and breaks the original application.
What Microsoft is doing is the right way to do it. Ideally, upon upgrade of a DLL, there should be a way for the application to contact the mother ship (i.e. the home page of the developer) and ask for approval to upgrade to the new DLL.
Wrong. Actually Xerox felt robbed, and sued Apple. Xerox lost in court because they took too long to bring forth the lawsuit, not because it did not have merit.
Slashdot Mirror
Yahoo denies that its new initiative is a declaration of war on Google. Eric E. Schmidt, 47, Google's chief executive, also says the two companies are still allies. But relations are strained.
Is he being diplomatic, or is just in denial. I wasn't that impressed with Eric's performance at the helm of Novel (IMHO the mere fact that he took the job showed incompetence). And why did Serge Brin choose Eric Schmidt as CEO? 'cuz it was the only candidate that had been to burning man'.
As Marc Andressen said, once you are big enough, you appear in the radar scopes of the big guys, and your true mettle is tested. So far the comments above from Eric Schmidt and Serge Brin give me little confidence....
- an index of even just 10% of the web would satisfy most people, as long as it is the right 10% and is searched effectively.
Altavista tried this, Lycos tried this, excite tried it too, and guess what? users did mind. They would get really pissed that they couldn't find a web page they had seen earlier.
Traditionally, the Economist has been one of the earliest media outlets to get technology. The first printed press reference to the Internet outside the tech press is from the Economist, ditto for Linux, but this time around they are way off the mark.
The reason Linux is so popular is not that is free. BeOS is also free. Linux is successful due to convergence of many different factors:
(1) Free
(2) Open source
(3) Unix compatible
(4) X-windows (X11) compatible
(5) designed for x86 (yes it runs on many other chips, still Linux is an x86 project from the get go)
(6) Multiple vendor supported
(7) Plenty of third party support
Moreover each of these things feed of each other. That is why Linux is so popular.
Sonny, listen to me. We geeks have a tendency to either be incredibly cheap or get so overly excited about finally having a girlfriend that we go overboard on gifts.
Giving her an expensive gift will not improve your image in her eyes. To the contrary, it will send the message "I'm so desperate I'm willing to shell $3K to impresss you". Knowing she has you wrapped around her finger will encourage her to search for another potential boyfriend, as she knows there is a fail-safe option at home.
Don't get me wrong. Be nice to your GF but don't over do it, and occasionally you might want to hold back a little on purpose so she yearns for you.
Save the $3k and promise to visit her in a few months. If she really loves you, the thought of having you in Japan (for a _visit_, you dirty minded folks) would be gift enough for her.
Do you know the history of C? Do you know why it was created?
/.ers have come to assume that C/Unix/whatever are written in stone and cannot possibly be improved (contrast this with perl apocalipsis and python 3, which are efforts starting from the assumption that lots of things could be fixed).
Of course. Algol and Pascal were introduced at the same time yet do not have this obnoxious behaviour. I bet you didn't know that.
You wouldn't want to accidently activate the whole set of floating-point division routines unless you really wanted to.
Let's assume for the sake of the argument that is true, why not fix it now? For example, sqrt(2) used to evaluate to garbage. The latest version of C does the proper casting to float and returns 1.41...
I really don't see why so many
People are so used to this design flaw that they can no longer tell that it is there. What would you say if I wrote a language in which the name of rebooting subroutine was say, "print"?
.sig is no different. It looks like standard schoolbook arithmetic, hence the outcome should be a resonable approximation of it.
Well the code in my
Yes, Java has the same flaw as it inherited several C/C++ design flaws. In fact the bug can be traced back to Fortran.
Pascal and the new version of Python, algol, modula and others don't have this flaw as they use div for integer division.
Good, safe, secure and efficient code is generated by educated programmers who are aware of what they're doing.
Ah the typical real men need no bounds check argument. This is, of course, a bogus argument.
In real life, people cannot be expected to be extra careful day in and day out. It's just not humanly possible. The long history of buffer overrun exploits proves this.
Coders make mistakes. That's why they put a backspace key on keyboards.
That's only there for PC wimps.
Everybody knows that pressing backspace in the original Emacs brought up the help page (I'm not making this up).
Seriously, with the kind of computational power that is now available even in a bottom-of-the-line gray-box PC the time has come to retire mainstream languages that do not have array bounds checking built-in.
<rant>
We are placing the entire computing infrastructure of the nation at risk in the name of a few measly clock cycles that would be spent performing bounds check.
Engineers know that the only way to have realiable systems is to have substantial safety margins and backup systems. This is a lesson that we software developers ignore at our own peril.
</rant>
However, IBM's contributions to Linux pale, or become transparent to be more accurate, in comparison to the dedicated hackers from all over the world who contribute to Linux.
If IBM went to the trouble of hiring companies to improve Linux then some development that they wanted to see through wasn't being done. How important it was? important enough for IBM to spend millions of dollars on this.
Keep im mind that some of the work that makes an OS entreprise strength is nitty gritty optimization that the hackers tend to skirt, just like they avoid quality GUI issues (designing tons of star trek inspired skins==cool, designing *one* usable skin for the entire Linux desktop==uncool).
While I have no idea if SCO has a case or not, I see that many here assume IBM had nothing to do with making Linux enterprise stable, and scoff at SCO's claim.
Yet, if you take the time to google the web you'll find that IBM dedicated an entire internal group to Linux and hired several external companies during 1999-2001 with the sole purpose of making Linux entreprise strength (even Linus has said so).
Now, to be clear, this does not yet prove that any illegal transfer of technology took place (and I doubt SCO will be able to prove it, IMHO they are fishing hoping to find the smoking gun during discovery), but it does verify one of the main three claims from SCO.
one million dollars!
....for once an extinction that wasn't really Microsoft's fault.
....for once a lawsuit that isn't really Microsoft's fault.
They have the resources to fix their security problems. They just haven't. Even when they know about a problem, it often goes unfixed for months, if not longer.
Sure but this is a different problem: Microsoft is unresponsive.
So let's not confuse quality of released code and responsiveness of manufacturer.
Personally, I agree with your comment on M$ unresponsiveness. We once found a bug in the optimization routine of Microsoft C. We had the platinum (or whatever the name is) support package with M$ which included 'face' time with "experts" so we sent an e-mail with the bug report.
Their answer: there won't be a patch forthcoming, it won't be fixed any time soon, it is not even clear if it will be fixed a year and half from now when we release the new version.
This is in contrast with other software manufacturers for which we would often get patches sent to us within a month or so of reporting a bug.
Ever hear anything good said about MFC? What about ATL? How about their OLE implementation?
Actually, Miguel de Icaza of gnome fame speaks highly of OLE... but that is besides my point.
I'd never argue that M$ puts out quality code. My point is that M$ code is crappy but no worse than that put out by competitors.
This is based on my use of several different platforms during the day (Unix, Linux, Apple, WinXP/2K) as well as software from many different manufacturers. Microsoft products do not stand out as being particularly more crash prone compared than the rest.
p.s. In fact currently the only piece of software that currently stands out as particularly buggy is my beta version of Mozilla 1.3b, which is quite understandable....
At the same time, from a security standpoint things are as bad as ever. Of all the machines on my network here, the only ones that have ever been compromised are the Windows boxes.
I agree that this is the case. Yet, a sensible explanation of why this is so is that more people write Windows exploits than Unix. Hence more flaws are found. The old adage of OSS is: debugging is fungible. Well so is cracking. Let's call that Alomex Law "cracking is fungible", as a consequence the most popular platform will be the most cracked. In fact as Linux continues to gain ground we'll see an increase in exploits (side note: about a year ago all our Red Hat boxes were cracked here).
In simple terms, if my goal as evil-cracker is to maximize disruption why would I spend hours pouring BeOS code that would give me access to four computers, when I can build upon readily available cracking utilities that can give me access to 90% of the world's computers?
It seems to me that one of the reasons this issue is so controversial is because it would mean we have to update our conception of reality, be it dropping Pluto or adding Ceres, and people just hate that.
I remember back in the 70s when telling people that Neptune was farther out (at that time) than Pluto, they would get very upset and push aside the SciAm diagram explaining the intersecting orbits. These are people who would otherwise be found a kid talking about the stars and trips to the moon quite endearing.
And ten years later at the next high school reunion, everyone will feel sorry for you since and your class mates (while poorer) had time to live their life (have family, friends, and kids) while you worked all your free time away at Microsoft.
Feel sorry for you?? Clearly you haven't been to your ten year school reunion.
They are the best ego trip ever for us geeks!
Turns out that the bimmer you are driving, which is non-distinct among the Benz kompressors and Porsche's carrera in your company parking lot is the best car of those in your class. It also turns out that you are making the most money. You are wearing the nicest suit and all the chicks that use to shun you drool over you. Odds are that you'll get some by the end of the night.
ymmv, asda.
So if it's such a great place to work, and they're woo'ing all of the best intern material, WHY do they continue to produce such crap code?!?
Do they really or is it just a urban myth?
People used to complain about how buggy IE was, but when finally Netscape's code was released for Mozilla turned out it was no better. So much so that Mozilla had to discard Netscape's code.
Another example, in 1994 I was routinely using a Mac, and it would crash ever one or two hours (cooperative multitasking anyone?). Back then you could read anywhere how bad an unstable Win 3.1 was, but you'd never hear a peep from the Apple camp.
1) get graduates straight out of universtiy.
Right out of university? Without even letting them join the unemployed lines for a while? The _gall_
2) condition them to believe working extremely long hours and weekends is "normal".
As compared to all other programming jobs in Sillicon Valley, where they want you there only from 9 to 5, and they send you home if you are there past 5.............NOT!
3) condition them to believe that if you're a real professional then your work is more important than socialising with your friends and spending time with your family.
Huh? M$ and other software companies only hire nerds. They come preconditioned to believe that. You don't believe me?
4) pay them relatively low salaries, but promise big ones in the future.
In your dreams. M$ used to pay below average salaries but give generous stock options, for a way above average salary. A couple of years back when the stock price stopped moving up, M$ instituted a pay increase across the board...
5) give them free pop, sweeties and toys.
This is supposed to be bad, or you are just jealous?
6) See how far you can push the suckers!
While all other companies only have your best interests in mind right?
No, it's wrong. You merely have the blind leading the blind.
Bullshit. Implementing a fix that is not QA'd and fully tested should never be done. Clearly you have no idea how to develop robust, reliable software.
The real solution is to indicate whether a DLL is a bug-fix release or whether it represents a significant and incompatible change to the APIs.
/* string terminated by \0\n\0 due to DLL bug */
This doesn't work, as often application developers detected the original bug, e.g.
then the fix gets released and breaks the original application.
What Microsoft is doing is the right way to do it.
Ideally, upon upgrade of a DLL, there should be a way for the application to contact the mother ship (i.e. the home page of the developer) and ask for approval to upgrade to the new DLL.
Wrong. Actually Xerox felt robbed, and sued Apple. Xerox lost in court because they took too long to bring forth the lawsuit, not because it did not have merit.