I totally agree with you. Structure and Interpretation Computer Programs is perhaps the best Computer Science and/or programming book I have ever read!
A lot of people have become slightly unhappy with IronPython. The development has been slow and private and the license is questionable.
Here is what Edd Dumbill, a Mono developer had to say about it:
Perhaps more seriously, third-party patches won't be considered until after the 1.0 release. Hugunin encourages people to be involved, but only in filing good bug reports and feature requests. And when doing this involves a.NET passport, and using the GotDotNet web forums rather than good old mailing lists, it's a bit of a disincentive.
Added to that, there's some uncertainty about the freeness of IronPython's license. While it looks free, it's got the same name, "Shared Source", as several Microsoft licenses that definitely are not free.
Does not sound good. Edd recommends for people looking for a Python experience on.Net to try boo which is similar to Python. I have read some Python blogs that think this is a bad idea but I think is because they are so Python-centric and the suggested solution is not called Python. Oh well, another tear for no one.
The title Plug-ins and Extensible Architectures made me thing of Lisp and Emacs.
Kinda funny it turned out to be about Eclipse and by association Java.
The sentence: Developers who want a flexible, configurable, IDE have long preferred plug-in architectures such as Eclipse over what they might view as the bloated, monolithic alternatives. Made me kinda scratch my head, Emacs has been around longer and is arguably more extensible.
As for Eclipse I am reminded of Greenspun's Tenth Rule of Programming: "Any sufficiently complicated C or Fortran program contains an ad-hoc, informally-specified bug-ridden slow implementation of half of Common Lisp." I could butcher that quote and turn it into shitface's Law: "Any sufficiently complicated editor or IDE is an ad-hoc and informally-specified bug ridden implementation of half of Emacs."
If as you put it " Lisp has and always will have an ass syntax and crap performance" then why have you been using it for the past twenty years?
The point that the other poster made about m-expressions is also very valid- there is nothing stopping you, or anyone else, from creating a new syntax for Lixp.
As for your speed concerns, in my experience Lisp is faster than Java (but what isn't). And good Lisp systems have been shown to generate code better C compilers.
Is it really a good job to judge a language from a tutorial? Most old C++ books were just bloated C tutorials with the addition of references and a primitive sense of classes (used to group functions)- I don't think too many people would call that C++ (and we are talking just 5 years ago- so not that long ago). By the way, I found the best book for learning C++ is The C++ Programming Language, 3rd Edition by Bjarne Stroustrup- it might be advanced but nothing else really teaches C++.
As for your loop issues how about: (loop for i from 0 to 10 do (format t "~A~%" i))? Common Lisp has a lot of looping macros available. Common Lisp does have goto- called go but its use is mostly frowned upon. Most paradigms you can think of you really will find in Common Lisp.
Ugh, too much thinking like that is exactly what is wrong with slashdot. The man answered the question the way he wanted to and that is it- live with it. Removing his answer now would make the threads concerning his answer even less relevant. You and everybody else that don't like his answer need to just shut up- this is the United States of America and the man can think say as he wishes.
I do not know anything about this hello im but it does not sound to be using the jabber protocol which google was rumored to be pursuing before. Anyone know anything about hello?
What kind of cars would we have today if everyone had just copied Ford's Model T? One cannot just copy the competition because people are familiar with it. Copying the competition is bad if for no other reason because you are letting them define the market. There is plenty room for improvement in the ui arena.
There has to be a reason why mac people don't jump ship to another desktop and at the same time so many people from other desktops, primarily Windows people, switch to macs. It should be noted:
the Mac UI is different than Windows and pretty much everything else
until recently the Mac was not exactly crash-free since it lacked such things as memory protection
the Mac ui certainly is not free
I am not suggesting that we copy the Mac UI, but I think it provides proof that there is something better out there.
Copying Microsoft will not attract your average computer user- they are either happy enough with Windows or unwilling/unknowledgeable to try Linux (GNOME or KDE). In the past 10 years, what was the single most thing that Microsoft did that was attributed to attracting new users? Ditch their old Windows 3.1 interface and adopt the "more mac-like" Windows 95 interface. As far as UI's go it was a big leap forward in the Windows community. Why should we be playing catch-up when we could be making the next big leap?
Of course, there is the propaganda or spin to consider also. The average computer user believes the big leaper in the industry is Microsoft despite the fact that any serious ui theory I know of has only come out of Apple and before them Xerox. Microsoft is able to pretend to be this big leaper because instead of trying to prove the strengths of their UI they dismiss Apple's with the Macs cost a lot rhetoric that you were babbling. Why copy pretenders?
The same but better is contradictory and not good enough for me. Perhaps, similar but better? I want just plain better.
As far as your (writing) style goes, I am (almost) sure that it could be improved upon. I found your (over-) use of parenthesis offensive. I prefer "it is open" instead of "it is (almost) free" for example.
Well far as the speed of method calls go there is a need for it in Java because of the lack of compiler inlining. A good C++ program can be agressively inlined to a point that function calling overhead is not an issue. Inlining is a big part of the reason why C++'s sort regularly beats C's qsort (there are some algorithmic issues also). Yes, JVM's usually do jit compiling which includes inlining but there is a cost for the jitting and is it me or is there a lot more method calls in Java.
As I have said before, the Fibonacci "benchmark" is horrible. I can write a C++ program that makes use of the recursive definition and yet spend no time computing the values at runtime. Real programs are never as simple as compute 3000 fibonacci numbers and exit. Most importantly, there is a closed formed to computing Fibonacci numbers so why use the naive recursive definition? Did you get the performance increase with the closed formed?
I am not going to get sucked into saying that Java is slow or that it is faster than C and C++ or assembly or whatever else you like to claim or dispute. Java is what it is.
With all that said I really doubt it is the C hackers and people citing java device drivers as the reason that Java has a reputation for being slow.
I have always felt that Java is percieved as being slow because of the programmers it attracts. Namely beginners and underskilled programmers because Java is believed to be friendly and forgiving for such. Combine that with the explosion of open source and the pc revolution and you have a lot projects seeing daylight that have no business doing so. Yes there are talented Java programmers out there but there are a lot more mediocre Java programmers out there (this is true of any language community but it seems more relevant when talking about Java, perhaps because Java is(/was?) in demand so a lot of people are claiming to know it when they have no right doing so).
As for the device driver argument(?), the programmer in me says one should be able to write a (fast) device driver in Java (or just about any language). Java does not make it easy because it is difficult to get to the needed low levelness with Java required for a device driver but that can be worked around. There are two things that are holding back java device drivers the developers and the compilers. It would take a special programmer to write a really decent device driver. As for the compilers you are talking native code generation, the device driver writer most likely having initimate knowledge of the compiler, and possibly dropping garbage collection. Why enable the garbage collector if you know you do not need it? BTW, obviously I am talking about native device drivers and not that JINI stuff.
Perhaps I am nitpicking but the name IronPython gives me the false impression of enabling Python to run on bare metal. The name would be much better suited for a Python system that compiled down to native code, sigh.
I have been kicking around some Python system implementations. Basically, people like to compare Python to Lisp and there are some decent (better than the "native C Python implementation") Lisp systems out there. Connect the dots.
I am not a macros expert or anything but the examples of Scheme hygenic macros in the article are in bad taste if not just plain wrong. The macro examples look more like common lisp macros than the hygenic macros defined in the R5RS. The article talks about Scheme macros using quasi-quote but (R5RS) macros do not make use of quasi-quote for their expansion. With all that said most Scheme systems do provide such a macro system but I believe they are non-hygenic. I have even seen people on comp.lang.scheme more or less get frustrated with the hygenic macros and roll their own macro system with quasi-quote and eval.
Knowing this is wrong I am suspecious of the entire article now. Given I think such articles are more important for the ideas within them and not so much their re-hashing of history but it to get something wrong still hurts.
Reminds me of some SNL skit for some fake company that comes and cleans up your personal space right after your death so no one finds those porno magazines and other stuff that you would not want your family finding. It ended by saying something like "You might be going to hell but there is no reason that your loved ones should know it" or something like that.
On a serious note, of course some personal websites live on like Rich Stevens site. If you it is an important site then it would be continued but how many useful personal sites are there? Ever go to Linus's personal site- it weak in content and meaningless to anyone that uses Linux. I think most personal sites are more like Linus's, weak in content and meaningless to whatever the person really does, but of course their are some useful personal sites like Stevens. With blogs and what not I guess this is becoming more of any issue. Every time I read some blogs at ximian I see a link to Ettore's and don't know what to feel.
I never heard of it until last summer and I am horrible at it but it is great. My friends and I would gather in some guys basement and play for hours. It was great fun laughing at how funny everybody looked trying to dance. One of my friends would keep losing his balance and would just grab the ceiling to cope. The guy whose basement it was always dissappointed that we seemed more interesting in making fun of each other instead of encouraging each other to improve but it was fun. It is one of those games you play in a social setting and don't so much care how good you are at it.
I know this is off topic but I doubt TUTOS will ever get any significant mention here . . . so this might my best chance to rant against it. A few semesters ago I had a professor who decided that he wanted a Microsoft Project clone and somehow decided that TUTOS was a good start. Right away we decided that TUTOS's interface was screwed up beyond repair and its set of features just seemed thrown together out. So almost right away we were committed to writing a completely new interface, in java, and refining the features. Two or three weeks into the probject the only thing we had in common with TUTOS was the database and *needed* to add additional tables there to enforce task dependencies. I realize that TUTOS is not project management software but rather team organization software (wtf is that- is that some term they invented?) but it just seemed like a really big calendar system to us. Needless to say the class never finished the project. Our failure , of course, was not completely to blame on TUTOS but also just bad management and really demanding time constraints. It is funny how that is the only project I have every worked on and I already feel that I could make a useful contribution to the huge pile of software engineering books out there.
I think I am getting all teared up. I remember back in the day when I trolled here. Now I am semi-legit. Its no fun to troll now- I guess Taco achieved what he wanted. Me and my friends got into a arguement one time about whether Cmdr stands for Commander or Comodore- what is your opinion?
This really is not news. I can tell you where it all starts though: Microsoft. Microsoft stopped supplying CD's to the OEM's to pass along to the customer citing it was the OEM's responsibility to make a recovery CD and supply the customer. So now you have OEM's paying that infamous Microsoft tax but the customer has even less to show for it now- think Microsoft is a monopoly? I guess HP, like a lot of other OEM's, don't want to put in that little extra effort to do things right.
You are missing the point. QT is a toolkit which is used in essentially every kde application. Yes QT has been available in some sort of free form but the problem was what if some day that ceased to be the case? Now ximian is selling one little plugin for use with their open source product and some dickhead is bitching about it. Evolution certainly does not depend on this plugin but KDE depends very much on QT. Besides I think it makes sense to sell this plugin. Ximian has to make money somehow and this seems fair. Ximian does not even compromise their open source beliefs because the basic product is still free software. The basic software for KDE has not always been free software and that was the problem.
I don't how to put this nicely so here we go. The current "stable" linux seems to be in name only. I am hearing horror stories of people turning off dma so that their fs don't get ruined and then there is the whole vm situation. How do you plan on stablizing 2.4? Are you gonna be the type the fixes these problems himself or are you gonna be more passive and wait for patches?
>> face it C++ is not a great OO language and gtk is not shining example of C++
You are right about gtk not being a shining example of C++ but only because gtk is written in C!
>> KDE definitely was wise to use Qt and made the correct decision
Okay, you find it fit to bash C++ but at the same time you are praising QT- which is a C++ toolkit/framework! Do you have any idea what you are talking about?
>> I say Gnome should dump GTK and C++ with it's half baked simula object model, multiple inheritance, clumsy object design, and it's fragile base class issue a major problem in big development.
Umm, besides the gtk thing you just described KDE and in the right hands C++ is not as bad you make it. Once again I must question if you know what you are talking about.
>> Long live C and hopefully C++ will see it use in places where it is appropriate as opposed to today where it is used everywhere.
For better or worse, that sums up what GNOME is about.
Your whole general notion that GNOME is unsavable is stupid! Well, besides you knowing nothing about the codebases that you are trying to talk about you must understand that no project is open source project is unsavable! KDE 2.0 is basically a complete re-write from KDE 1.x- what prevents people from re-writing bad code/libraries? Given in the commercial world- a complete re-write could be doom but the open source world is not bounded by such finicial concerns and the code is available for anyone to change/improve. If you really think that GNOME is in as bad of shape as you claim then fix it!
bubble sorting partial list from quicksort? yeah right. the rule is you only use bubble sort on tests (because it is the easiest to write). If you want a simple sort algorithm for sorting partials you best bet is insertion sort followed by selection. You are an idiot.
MSN is obviously not a real news (or portal) site but just an off-shoot of microsoft that is aims to extend microsofts stronghold. Why is it that microsoft can create a very viewable website for microsoft.com (where they pedal their wares) but can't do so with msn.com? The news does not have to be fancy just informative- is there anything fancy about yahoo or even cnn? This is stupid microsoft tried to make a very small percentage of people start to use internet explorer not because it is better but just because microsoft can exercise that kind of power. Remember: "power corrupts and absolute power corrupts absolutely!"
Slashdot Mirror
Right now I am reading Compilers: Principles, Techniques, and Tools by Aho, Sethi, and Ullman. Great book.
I understand there is a "new" edition coming out; I think the new addition will be called 21st Century Compilers.
I totally agree with you. Structure and Interpretation Computer Programs is perhaps the best Computer Science and/or programming book I have ever read!
Hope more people take note of your comment.
Here is what Edd Dumbill, a Mono developer had to say about it:
Does not sound good. Edd recommends for people looking for a Python experience on
The title Plug-ins and Extensible Architectures made me thing of Lisp and Emacs.
Kinda funny it turned out to be about Eclipse and by association Java.
The sentence: Developers who want a flexible, configurable, IDE have long preferred plug-in architectures such as Eclipse over what they might view as the bloated, monolithic alternatives. Made me kinda scratch my head, Emacs has been around longer and is arguably more extensible.
As for Eclipse I am reminded of Greenspun's Tenth Rule of Programming: "Any sufficiently complicated C or Fortran program contains an ad-hoc, informally-specified bug-ridden slow implementation of half of Common Lisp." I could butcher that quote and turn it into shitface's Law: "Any sufficiently complicated editor or IDE is an ad-hoc and informally-specified bug ridden implementation of half of Emacs."
You are starting to sound a lot like some guy who knows a lot of languages but is a master of none of them.
If as you put it " Lisp has and always will have an ass syntax and crap performance" then why have you been using it for the past twenty years?
The point that the other poster made about m-expressions is also very valid- there is nothing stopping you, or anyone else, from creating a new syntax for Lixp.
As for your speed concerns, in my experience Lisp is faster than Java (but what isn't). And good Lisp systems have been shown to generate code better C compilers.
Is it really a good job to judge a language from a tutorial? Most old C++ books were just bloated C tutorials with the addition of references and a primitive sense of classes (used to group functions)- I don't think too many people would call that C++ (and we are talking just 5 years ago- so not that long ago). By the way, I found the best book for learning C++ is The C++ Programming Language, 3rd Edition by Bjarne Stroustrup- it might be advanced but nothing else really teaches C++.
As for your loop issues how about: (loop for i from 0 to 10 do (format t "~A~%" i))? Common Lisp has a lot of looping macros available. Common Lisp does have goto- called go but its use is mostly frowned upon. Most paradigms you can think of you really will find in Common Lisp.
Ugh, too much thinking like that is exactly what is wrong with slashdot. The man answered the question the way he wanted to and that is it- live with it. Removing his answer now would make the threads concerning his answer even less relevant. You and everybody else that don't like his answer need to just shut up- this is the United States of America and the man can think say as he wishes.
I do not know anything about this hello im but it does not sound to be using the jabber protocol which google was rumored to be pursuing before. Anyone know anything about hello?
There has to be a reason why mac people don't jump ship to another desktop and at the same time so many people from other desktops, primarily Windows people, switch to macs. It should be noted:
I am not suggesting that we copy the Mac UI, but I think it provides proof that there is something better out there.
Copying Microsoft will not attract your average computer user- they are either happy enough with Windows or unwilling/unknowledgeable to try Linux (GNOME or KDE). In the past 10 years, what was the single most thing that Microsoft did that was attributed to attracting new users? Ditch their old Windows 3.1 interface and adopt the "more mac-like" Windows 95 interface. As far as UI's go it was a big leap forward in the Windows community. Why should we be playing catch-up when we could be making the next big leap?
Of course, there is the propaganda or spin to consider also. The average computer user believes the big leaper in the industry is Microsoft despite the fact that any serious ui theory I know of has only come out of Apple and before them Xerox. Microsoft is able to pretend to be this big leaper because instead of trying to prove the strengths of their UI they dismiss Apple's with the Macs cost a lot rhetoric that you were babbling. Why copy pretenders?
The same but better is contradictory and not good enough for me. Perhaps, similar but better? I want just plain better.
As far as your (writing) style goes, I am (almost) sure that it could be improved upon. I found your (over-) use of parenthesis offensive. I prefer "it is open" instead of "it is (almost) free" for example.
Well far as the speed of method calls go there is a need for it in Java because of the lack of compiler inlining. A good C++ program can be agressively inlined to a point that function calling overhead is not an issue. Inlining is a big part of the reason why C++'s sort regularly beats C's qsort (there are some algorithmic issues also). Yes, JVM's usually do jit compiling which includes inlining but there is a cost for the jitting and is it me or is there a lot more method calls in Java.
As I have said before, the Fibonacci "benchmark" is horrible. I can write a C++ program that makes use of the recursive definition and yet spend no time computing the values at runtime. Real programs are never as simple as compute 3000 fibonacci numbers and exit. Most importantly, there is a closed formed to computing Fibonacci numbers so why use the naive recursive definition? Did you get the performance increase with the closed formed?
I am not going to get sucked into saying that Java is slow or that it is faster than C and C++ or assembly or whatever else you like to claim or dispute. Java is what it is.
With all that said I really doubt it is the C hackers and people citing java device drivers as the reason that Java has a reputation for being slow.
I have always felt that Java is percieved as being slow because of the programmers it attracts. Namely beginners and underskilled programmers because Java is believed to be friendly and forgiving for such. Combine that with the explosion of open source and the pc revolution and you have a lot projects seeing daylight that have no business doing so. Yes there are talented Java programmers out there but there are a lot more mediocre Java programmers out there (this is true of any language community but it seems more relevant when talking about Java, perhaps because Java is(/was?) in demand so a lot of people are claiming to know it when they have no right doing so).
As for the device driver argument(?), the programmer in me says one should be able to write a (fast) device driver in Java (or just about any language). Java does not make it easy because it is difficult to get to the needed low levelness with Java required for a device driver but that can be worked around. There are two things that are holding back java device drivers the developers and the compilers. It would take a special programmer to write a really decent device driver. As for the compilers you are talking native code generation, the device driver writer most likely having initimate knowledge of the compiler, and possibly dropping garbage collection. Why enable the garbage collector if you know you do not need it? BTW, obviously I am talking about native device drivers and not that JINI stuff.
Perhaps I am nitpicking but the name IronPython gives me the false impression of enabling Python to run on bare metal. The name would be much better suited for a Python system that compiled down to native code, sigh.
I have been kicking around some Python system implementations. Basically, people like to compare Python to Lisp and there are some decent (better than the "native C Python implementation") Lisp systems out there. Connect the dots.
Peace out.
I am not a macros expert or anything but the examples of Scheme hygenic macros in the article are in bad taste if not just plain wrong. The macro examples look more like common lisp macros than the hygenic macros defined in the R5RS. The article talks about Scheme macros using quasi-quote but (R5RS) macros do not make use of quasi-quote for their expansion. With all that said most Scheme systems do provide such a macro system but I believe they are non-hygenic. I have even seen people on comp.lang.scheme more or less get frustrated with the hygenic macros and roll their own macro system with quasi-quote and eval.
Knowing this is wrong I am suspecious of the entire article now. Given I think such articles are more important for the ideas within them and not so much their re-hashing of history but it to get something wrong still hurts.
Reminds me of some SNL skit for some fake company that comes and cleans up your personal space right after your death so no one finds those porno magazines and other stuff that you would not want your family finding. It ended by saying something like "You might be going to hell but there is no reason that your loved ones should know it" or something like that.
On a serious note, of course some personal websites live on like Rich Stevens site. If you it is an important site then it would be continued but how many useful personal sites are there? Ever go to Linus's personal site- it weak in content and meaningless to anyone that uses Linux. I think most personal sites are more like Linus's, weak in content and meaningless to whatever the person really does, but of course their are some useful personal sites like Stevens. With blogs and what not I guess this is becoming more of any issue. Every time I read some blogs at ximian I see a link to Ettore's and don't know what to feel.
I never heard of it until last summer and I am horrible at it but it is great. My friends and I would gather in some guys basement and play for hours. It was great fun laughing at how funny everybody looked trying to dance. One of my friends would keep losing his balance and would just grab the ceiling to cope. The guy whose basement it was always dissappointed that we seemed more interesting in making fun of each other instead of encouraging each other to improve but it was fun. It is one of those games you play in a social setting and don't so much care how good you are at it.
I know this is off topic but I doubt TUTOS will ever get any significant mention here . . . so this might my best chance to rant against it. A few semesters ago I had a professor who decided that he wanted a Microsoft Project clone and somehow decided that TUTOS was a good start. Right away we decided that TUTOS's interface was screwed up beyond repair and its set of features just seemed thrown together out. So almost right away we were committed to writing a completely new interface, in java, and refining the features. Two or three weeks into the probject the only thing we had in common with TUTOS was the database and *needed* to add additional tables there to enforce task dependencies. I realize that TUTOS is not project management software but rather team organization software (wtf is that- is that some term they invented?) but it just seemed like a really big calendar system to us. Needless to say the class never finished the project. Our failure , of course, was not completely to blame on TUTOS but also just bad management and really demanding time constraints. It is funny how that is the only project I have every worked on and I already feel that I could make a useful contribution to the huge pile of software engineering books out there.
Any good programmer rarely has problems with indexing arrays out of bounds.
I think I am getting all teared up. I remember back in the day when I trolled here. Now I am semi-legit. Its no fun to troll now- I guess Taco achieved what he wanted. Me and my friends got into a arguement one time about whether Cmdr stands for Commander or Comodore- what is your opinion?
This really is not news. I can tell you where it all starts though: Microsoft. Microsoft stopped supplying CD's to the OEM's to pass along to the customer citing it was the OEM's responsibility to make a recovery CD and supply the customer. So now you have OEM's paying that infamous Microsoft tax but the customer has even less to show for it now- think Microsoft is a monopoly? I guess HP, like a lot of other OEM's, don't want to put in that little extra effort to do things right.
You are missing the point. QT is a toolkit which is used in essentially every kde application. Yes QT has been available in some sort of free form but the problem was what if some day that ceased to be the case? Now ximian is selling one little plugin for use with their open source product and some dickhead is bitching about it. Evolution certainly does not depend on this plugin but KDE depends very much on QT. Besides I think it makes sense to sell this plugin. Ximian has to make money somehow and this seems fair. Ximian does not even compromise their open source beliefs because the basic product is still free software. The basic software for KDE has not always been free software and that was the problem.
I don't how to put this nicely so here we go. The current "stable" linux seems to be in name only. I am hearing horror stories of people turning off dma so that their fs don't get ruined and then there is the whole vm situation. How do you plan on stablizing 2.4? Are you gonna be the type the fixes these problems himself or are you gonna be more passive and wait for patches?
You my friend are an idiot and a troll.
>> face it C++ is not a great OO language and gtk is not shining example of C++
You are right about gtk not being a shining example of C++ but only because gtk is written in C!
>> KDE definitely was wise to use Qt and made the correct decision
Okay, you find it fit to bash C++ but at the same time you are praising QT- which is a C++ toolkit/framework! Do you have any idea what you are talking about?
>> I say Gnome should dump GTK and C++ with it's half baked simula object model, multiple inheritance, clumsy object design, and it's fragile base class issue a major problem in big development.
Umm, besides the gtk thing you just described KDE and in the right hands C++ is not as bad you make it. Once again I must question if you know what you are talking about.
>> Long live C and hopefully C++ will see it use in places where it is appropriate as opposed to today where it is used everywhere.
For better or worse, that sums up what GNOME is about.
Your whole general notion that GNOME is unsavable is stupid! Well, besides you knowing nothing about the codebases that you are trying to talk about you must understand that no project is open source project is unsavable! KDE 2.0 is basically a complete re-write from KDE 1.x- what prevents people from re-writing bad code/libraries? Given in the commercial world- a complete re-write could be doom but the open source world is not bounded by such finicial concerns and the code is available for anyone to change/improve. If you really think that GNOME is in as bad of shape as you claim then fix it!
bubble sorting partial list from quicksort? yeah right. the rule is you only use bubble sort on tests (because it is the easiest to write). If you want a simple sort algorithm for sorting partials you best bet is insertion sort followed by selection. You are an idiot.
MSN is obviously not a real news (or portal) site but just an off-shoot of microsoft that is aims to extend microsofts stronghold. Why is it that microsoft can create a very viewable website for microsoft.com (where they pedal their wares) but can't do so with msn.com? The news does not have to be fancy just informative- is there anything fancy about yahoo or even cnn? This is stupid microsoft tried to make a very small percentage of people start to use internet explorer not because it is better but just because microsoft can exercise that kind of power. Remember: "power corrupts and absolute power corrupts absolutely!"