Linux-Advocacy Digest #497, Volume #29 Fri, 6 Oct 00 21:13:08 EDT
Contents:
Re: Why should anyone prefer Linux to Win2k on the DeskTop (Steve Mading)
Re: What kind of WinTroll Idiot are you anyway? (Steve Mading)
Re: What kind of WinTroll Idiot are you anyway? (Michael Marion)
Re: Do Linux suXX??? (Steve Mading)
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (Bob Hauck)
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (Richard)
Re: Why is MS copying Sun??? (Dennis Yelle)
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (FM)
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (Richard)
Re: Unix rules in Redmond ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: Steve Mading <[EMAIL PROTECTED]>
Subject: Re: Why should anyone prefer Linux to Win2k on the DeskTop
Date: 7 Oct 2000 00:15:47 GMT
Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
: "Steve Mading" <[EMAIL PROTECTED]> wrote in message
: news:8rj4si$mvq$[EMAIL PROTECTED]...
:> Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
:>
:> : And what makes you think that apache is used more widely? I don't think
:> : there are any statistics to back this up, unless of course you mean
: "More
:> : domain names are hosted on apache servers",
:>
:> Netcraft goes down to the level of individual hostnames, not domain
:> names. (so, for example, thishost.somebiz.com and thathost.somebiz.com
:> get counted as two hits even though they are the same domain name.)
:> Are you trying to imply that if an IIS site uses multiple machines
:> masquerading behind one hostname to serve multiple requests that
:> this should be counted as multiple IIS servers? Well, if that kind
:> of metric would help IIS becuase IIS is in that sort of situation
:> more often than Apache, then I put forth that this makes IIS look
:> bad, not good. Why all the extra servers to run one site? What's
:> the matter? One machine can't handle it?
: You are correct, I meant hostname, not domain name. My point was that we
: don't know for sure how many of those hostnames are running on the same
: server. If they point to the same IP address, it could be figured out, but
: that's not necessarily the case. You can have multiple IP's on a single NIC
: (or multiple NICs). Netcraft doesn't even try to filter out unique IP's.
: I'm sorry though, one machine simply cannot handle the loads of large sites
: which can move Terrabytes a day (Microsoft.com claims to transfer more than
: 6 terrabytes a day with over 4.1 million unique users a day)
: You really think one machine can handle that load? Two machines? 10?
: (remember, we're not just talking about transfering pages, but dynamic
: content, SQL server queries, site searches, etc..)
Certainly, given a high enough network load, ANY variety of
webserver, be it Apache/linux or IIS/NT, or whatever, will have
to split out to several servers. The point is, if IIS is set up
this way more often than apache, and *isn't* handling more traffic,
then IIS is handling less traffic per server. So your point
originally, about multiple servers hiding behind one hostname, cannot
possibly make IIS look better. You're right that the netcraft
survey is sloppy, but you are wrong in assuming that correcting
the slop in the case of hiding servers behind one hostname would
make IIS look better, for if it increases the IIS number of sites
to add that metric in, then that must mean IIS can't handle the
load with the same number of servers. The only way out of this
would be if IIS tends to serve sites with higher traffic so that
the need for these parallel servers would be justified. But,
as I'll get to down below, that's not the case.
: Well, I can't think of a site that has more traffic that MS. And that's not
: even counting Hotmail and bCentral. IBM's web site runs on Domino. Yahoo
: runs on some unknown web server running on FreeBSD. CNet runs on Netscape's
: server (as of course does Netscape).
: Which top 100 sites, as measured by traffic run Apache again? I can't think
: of many. About the only high traffic sites I know of running Apache are
: places like Slashdot, and they wouldn't even register on the top 100.
Look here and be shown: http://www.durak.org/sean/pubs/bss/
Here's a summary, using three of the lists of "top sites by traffic"
that exist:
Cross referencing the Alexa 1,000 list with server type gives this:
1st place - 459 sites on Apache (49.73%)
2nd place - 218 sites on Microsoft IIS (23.62%)
3rd place - 185 sites on Netscape (20.04%)
4th place - 16 sites on zeus ( 1.73%)
... and a bunch of others at < 1% after this ...
Here's the same thing with the Media Metrix 500 list:
1 187 37.78% Apache
2 135 27.27% Microsoft
3 124 25.05% Netscape
4 2 0.40% AOLserver/3.0
... and a bunch of others < 1% after this...
Here's the same thing with the 100 Hot list:
1 44 48.35% apache
2 20 21.98% microsoft iis
3 16 17.58% netscape
... and a bunch of others < 2 % after this...
------------------------------
From: Steve Mading <[EMAIL PROTECTED]>
Subject: Re: What kind of WinTroll Idiot are you anyway?
Date: 7 Oct 2000 00:19:32 GMT
Gary Hallock <[EMAIL PROTECTED]> wrote:
: Steve Mading wrote:
:> rpm is great, but the documentation in the man page is complete crap
:> and needs to be re-written. It is full of examples, but none of them
:> cover the most commonly needed use of the command: I have an rpm
:> file I have not installed yet and I want to see what files it's going
:> to install. (Like the "tar tvf" command for tar.) It turns out
:> that all the "query" commands for rpm assume you are trying to query
:> the already installed packages, but this isn't really stated very
:> clearly in the manpage.
: The -p option will look at an uninstalled rpm file:
: rpm -qlp rpmfile
: You should take a look at "Maxium RPM" downloadable from www.rpm.org
I wasn't saying that the option was nonexistant, just that the man
page does a piss-poor job of mentioning it, or the need for it.
(It is in there somewhere, but it's buried deep).
------------------------------
From: Michael Marion <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: What kind of WinTroll Idiot are you anyway?
Date: Sat, 07 Oct 2000 00:40:07 GMT
Nathaniel Jay Lee wrote:
> person that hasn't heard these complaints. I can't even
> use the 'official' NVIDIA driver under Linux because it's
> 'kernel-space enhancements' lock up X constantly (although
> why it doesn't cause a kernel dump I don't know). I won't
> lie and say I have any experience with NT or 9x with this
> card, but people that have 'actual' experience say they
> see some of the same sort of behavior under Windows. So
Yep.. I'm one that had issues under 2k (and even some in 98).. I tried to
explain to him that there were several hardware configurations tried, multiple
drivers, etc.. but he refuses to believe me. What's even sadder is that
people posted links to pages with many stories of issues with nvidia drivers..
and he still denies it.
I think nvidia's hardware rocks.. there drivers, though, have lacked a lot
recently (until Detonator 3 that is).
--
Mike Marion - Unix SysAdmin/Engineer, Qualcomm Inc. - http://www.miguelito.org
If you were building The Matrix: NT or Unix? I thought so :)
--another /.-er
------------------------------
From: Steve Mading <[EMAIL PROTECTED]>
Subject: Re: Do Linux suXX???
Date: 7 Oct 2000 00:35:01 GMT
John Sanders <[EMAIL PROTECTED]> wrote:
: Unsupported personal opinion. The main theme seems to be that Linux is
: destroying the field of programming. That seems pretty amazing for an
: OS that's not going anywhere, and written by people that take no
: personal responsibility for their programming because they are 'hackers'
: and not 'engineers'. How could Linux screw up the Computer Science
: field so quickly?
My complaint about that article is how it forgets one very important
fact - opensource software is released in alpha state because that's
how you get more people to help you work on it. Grabbing a package
which is *advertised* as still being in-progress, and then bitching
about it not being perfect is akin to running into some software
company's building, grabbing a copy of their code off the programmer's
platform, and then bitching that it isn't ready yet. The only reason
some opensource projects are being publicly dissemated is because
the programmers working on it don't all live in the same building
with their code locked up in the same vault. This is how they work
together.
------------------------------
From: [EMAIL PROTECTED] (Bob Hauck)
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Reply-To: bobh{at}haucks{dot}org
Date: Sat, 07 Oct 2000 00:43:42 GMT
On 6 Oct 2000 21:21:06 GMT, Donovan Rebbechi <[EMAIL PROTECTED]> wrote:
>In java, even the primitive types are objects.
No, they aren't. Java is like C++ in that respect. The library
includes wrappers for the primitive types to make what you're saying
true, but this isn't really a "built in".
>In C++, they aren't ( and they really don't need to be )
There are advantages both ways. Mostly, the C++ way is more efficient
at runtime, the other way is somewhat more flexible and the language
doesn't need as many special cases.
>Well I guess you've named one feature. The fact that blocks of code are
>objects does not in itself seem to
It is helpful if you want to be able to load code at runtime. To
really take advantage of it you probably need to have an interpreted
environment though. I'm coming around to the view that this is ok if
the interpreter has an easy interface to C for the stuff that really
needs to be fast.
>C++ supports the object oriented paradigm, even if it's not purely object
>oriented. It's certainly no more "procedural" than it is OO.
One of the biggest problems with C++ is this lack of focus. It tries
to do everything and support all paradigms. That's why it is so big
and one reason the syntax is so ugly (I think it is uglier than C).
>Java comes awfully close to "everything is an object", so even in this
>definition, java is OO. Calling java a "procedural language" is the
>height of ignorance. It's difficult to write procedural code in java.
Yup. Java is an OO language by any reasonable definition. So is C++,
but C++ is also a lot of other things. It seems to be about three
different languages all packed into one box.
--
-| Bob Hauck
-| To Whom You Are Speaking
-| http://www.haucks.org/
------------------------------
From: Richard <[EMAIL PROTECTED]>
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Date: Sat, 07 Oct 2000 00:50:15 GMT
Donovan Rebbechi wrote:
> In C++, you can make classes objects if you like.
Which means that C++ classes are *not* objects.
What you can do with a Turing complete language is fucking irrelevant.
It's like claiming that normal English is well-suited to some subdomain
because you can define completely new words that are well-suited to it!
Or even better, like claiming that English is beautiful because you can
write beautiful things in it.
> Read a book on
> design patterns some time. In java, even the primitive types are
> objects.
In Java, primitive types are NOT objects. You have objects that
contain primitive types and you have primitive types, don't
confuse the two.
> In C++, they aren't ( and they really don't need to be )
If you don't care whether or not C++ is OO then why do you say it is?
> > Objects in Smalltalk cannot
> >access each other's parts (in C++ friends can access each other's private
> >parts).
^^^^^^^^^^^^^^^^^^^^^^^^^^^
Damnit, that was supposed to be "play with each other's private parts".
> I don't see how that makes C++ "less object oriented". If you don't like
It makes C++ disgusting. It also DOES make C++ non-OO since people
understand objects to have certain properties like solidity and non-
interpenetrability which you have just admitted C++ does not have.
> friends, don't use them. C++ is a multi paradigm language. You can
> program OO in it if you like, but it doesn't force you to.
IOW, C++ is not OO. "supporting" OO is not the same thing as *being* OO.
> I don't see how this is a big issue. I don't bother remembering the finer
> details of operator precedence.
Whether or not you know them is irrelevant. What's relevant is that
it takes your attention away from *what matters* when you're programming;
the objects.
> > In C++ blocks
> >of code don't exist as objects (in Java they don't exist as simple objects).
>
> Well I guess you've named one feature.
Wrong. I've given yet another example of something that is NOT AN OBJECT
in C++. Blocks of code are right there with unresolved messages and
execution contexts as far as not existing as objects in C++.
> The fact that blocks of code are
> objects does not in itself seem to
?
> >OO means "everything's an Object"
>
> I think a lot of people more authoritative than you would disagree
And I just bet that all of those people would know nothing about Self,
Smalltalk or Eiffel. They would all be ignorant C++ or Java programmers.
Smalltalk started this little clique called OOP and C++ programmers
were too lazy to fulfill the membership requirements so they started
telling themselves that they were part of the clique because it was
the cool thing to be at the time. This is how all hype works; the
semiconductors people appropriated the word "nanomachines" to refer
merely to machines at the nanometer scale even though all of the
existing excitement about nano-machines was about molecular manu-
facturing. What a fucknig lie.
> with this. If you add something that is not an "object" to something
> that is "object oriented", then do you have something that's no longer
> object oriented ?
*Precisely*.
> C++ supports the object oriented paradigm, even if it's not purely object
> oriented. It's certainly no more "procedural" than it is OO.
It's a chimera, a frankenstein's monster, that is neither procedural
nor OO. This does not make it /both/ procedural and OO, it makes it
*neither*.
A programmer *can* create an OO program in C++ but the language itself
isn't OO, will never be OO, and this forces programmers who want to
remain OO to reimplement everything that is not an object in C++. When
programming something, I don't need this kind of shit.
> Java comes awfully close to "everything is an object", so even in this
No, it doesn't. *Smalltalk* comes awfully close but not quite there.
For example, blocks are not full closures in (any?) Smalltalk dialects.
This violates OO quite nicely. Messages are also not objects, they
only get converted to objects if they fail. Variables are not objects.
Even SMALLTALK isn't perfectly OO so don't give me this shit about
Java being close. In Java "methods" in a class can access the private
members of all objects that belong to that class. What the fuck is the
deal with that? The only reason I can think of for this is to pander
to people who are too lazy or just don't know when to use callbacks!
> definition, java is OO. Calling java a "procedural language" is the
> height of ignorance. It's difficult to write procedural code in java.
> ( you don't even have "functions" in java, only methods )
And the difference between the two is?
Java is neither procedural nor OO, it's hype-oriented.
> Well, apart from philosophical beauty, it's not clear that the former
> has any tangible advantages ( i'm not going to argue about whether
> philosophical beauty is a tangible advantage or not )
<rolleyes> Learn to (seriously) program in Smalltalk and *then*
you'll be in a position to say it, but you won't unless you have
no problem with lying. Programming in an OOPL is a completely
different experience from programming in a non-OOPL language. You
are vastly more productive in it, it's easier to program in an OO
manner, everything in the system is readable by newbies (who might
not be able to understand the non-OO "features" of C++) and objects
are simply *easier* to understand than non-objects.
I love functional programmnig and I think it's very neat but when
it comes down to it, FP thinking simply isn't natural to humans.
OO thinking is. That directly affects the learnability and even
the *understandability* of a language.
And finally, in a world where you don't have to read a line of anyone
else's code, it might be fine if other people aren't as OO as you are.
That world is not the Real World.
> You're the one who came into this group spewing hateful and bigoted comments
> about "programmers". I believe most of the people here have decided which one
> of us "needs to grow up
And yet, I'm not the one who expects everyone else to take care of me.
> On the other hand, you, having made bold claims about being a "high level
> designer" seem unwilling or unable to address questions regarding your
> credibility. You've attacked me for calling you a kook, but you haven't
> refuted the claim.
I'm not even sure what the word means exactly ....
------------------------------
From: Dennis Yelle <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.arch,comp.arch
Subject: Re: Why is MS copying Sun???
Date: Fri, 06 Oct 2000 17:55:58 -0700
Paul Wallich wrote:
[...]
> According to the DR folks I talked to at the time, that wasn't quite
> accurate either. There was also the matter of an old-style IBM NDA
> that Big Blue's lawyers wanted Kildall to sign (along the lines of "you
> agree to an irrevocable option on your firstborn child before we even
> tell you what our names are"). I'm sure that story is a revised version as
> well, but when you have a monopoly position, you tend to be less willing
> to negotiate. Since MS had much less to lose from signing away its rights
> (and subsequently "reclaimed" them in questionable fashion), they were
> willing to go along with IBM.
In what sense did MS reclaim those rights?
Where can I find out more about this?
Dennis Yelle
--
I am a computer programmer and I am looking for a job.
There is a link to my resume here:
http://table.jps.net/~vert/
------------------------------
From: [EMAIL PROTECTED] (FM)
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Date: 7 Oct 2000 00:25:29 GMT
Reply-To: [EMAIL PROTECTED]
T. Max Devlin <[EMAIL PROTECTED]> wrote:
>Said FM in comp.os.linux.advocacy;
>>T. Max Devlin <[EMAIL PROTECTED]> wrote:
>>>You're a post-modernist imbecile.
>>Hehe. That's funny. It's the same guy who tried to argue that
>>there's an absolute standard for beauty and elegance, remember?
>No, I don't remember ever arguing such a thing. Or were you referring
>to Richard, who my comment was directed at.
Well yeah, "this guy" refers to Richard and "remember?" was
mostly a question towards you.
dan.
------------------------------
From: Richard <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.java.advocacy
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Date: Sat, 07 Oct 2000 00:59:14 GMT
The Ghost In The Machine wrote:
> comp.lang.java.advocacy added to newsgroups.
> It's far from clear whether C++'s methodology -- also known as
> "well, OK, I can be an object if I feel like it, but I don't have
> to be" -- is the best or not, although it works.
For some definition of 'works', COBOL works. :-)
> Java has an interesting dichotomy, and I'm still trying to figure
> out whether the following will work, on a theoretical level.
[I feel dizzy]
> [rest snipped, as I know little about Smalltalk -- which is a pity :-) ]
Amen, brother!
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Unix rules in Redmond
Date: Sat, 7 Oct 2000 02:22:26 +0200
Reply-To: [EMAIL PROTECTED]
In article <8rlb6h$ko2$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (.) wrote:
>
> Aw. Dresden doesnt know what a context switch is.
With a name like dressed in black and his porn connections I'm
sure he knows what a context switch is. It has nothing to do
with computing though. :-)
> Thats sweet.
How apropriate.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
