Linux-Advocacy Digest #268, Volume #34 Sun, 6 May 01 19:13:02 EDT
Contents:
Re: Justice Department LOVES Microsoft! (Rick)
Re: Justice Department LOVES Microsoft! (Rick)
Re: Justice Department LOVES Microsoft! ("Quantum Leaper")
Re: Justice Department LOVES Microsoft! (Rick)
Re: How to hack with a crash, another Microsoft "feature" ("Erik Funkenbusch")
Re: Article: AOL in cahoots with Compaq, HP to derail WinXP, .NET? (Nigel Feltham)
Re: The long slow slide to Microsoft.NOT (Giuliano Colla)
Re: Article: Want Media Player 8? Buy Windows XP ("Ayende Rahien")
Re: Richard Stallman what a tosser, and lies about free software ("Ayende Rahien")
Re: Why is Microsoft opening more Windows source code? ("Erik Funkenbusch")
Re: The _one_ thing that pisses me off about Linux (Richard Thrippleton)
----------------------------------------------------------------------------
From: Rick <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Justice Department LOVES Microsoft!
Date: Sun, 06 May 2001 18:11:23 -0400
Daniel Johnson wrote:
>
> "Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Daniel Johnson wrote:
> > > Then I suggest you are being sloppy with your accusations; you
> > > know full well that MS never excluded anyone. At their *worst*
> > > they want you to sell *their* product, whatever else you may sell.
> >
> > Then why were their OEM-licenses dependant upon NOT selling any
> > other vendor's OS?
>
> They weren't.
>
> Remember, not everything Max says is true. :D
Then, try expalain what a per-processor license is.
--
Rick
------------------------------
From: Rick <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Justice Department LOVES Microsoft!
Date: Sun, 06 May 2001 18:12:49 -0400
JS PL wrote:
>
> "T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Said Daniel Johnson in comp.os.linux.advocacy on Sat, 05 May 2001
> > 20:53:03 GMT;
> > >"Rick" <[EMAIL PROTECTED]> wrote in message
> > >news:[EMAIL PROTECTED]...
> > >> Daniel Johnson wrote:
> > >> > MS-DOS and DR-DOS were *both* lousy things
> > >> > to saddle Windows with. I'm not endorsing MS-DOS
> > >> > over DR-DOS; I'm endorsing Windows with as little
> > >> > of either as can be managed.
> > >>
> > >> Windows COULD NOT run without some DOS underneath. DR-DOS was superior.
> > >> And Microsoft used it monoply power to push DR-DOS out of the market.
> > >
> > >I don't see that DR-DOS was superior as a platform
> > >for Windows.
> >
> > Nice squirming, troll-boy.
> >
> > >The real trick was to supercede as much of DOS
> > >as possible, not to use another DOS.
> >
> > No wonder you find it so easy to be clueless; you think there's a
> > difference between something that is like DOS, and something that is
> > "another DOS". Its a wonder you can even find the power switch.
>
> Come on now!! Your pissing Max off!! The truth to him is like holy water on
> a vampire. Max thinks - Microsoft was supposed to create Windows to be
> compatible with DR DOS. He probably goes further into fantasyland to think
> Microsoft Inc. was supposed provide support for every Tom, Dick and DOS
> lookalike that came along.
NO, but then they should have put code into Windows to specifcally check
if it was running on another DOS, then throw up error messages to scare
off customers.
--
Rick
------------------------------
From: "Quantum Leaper" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Justice Department LOVES Microsoft!
Date: Sun, 06 May 2001 22:12:37 GMT
"Daniel Johnson" <[EMAIL PROTECTED]> wrote in message
news:3DXI6.6016$[EMAIL PROTECTED]...
> "Quantum Leaper" <[EMAIL PROTECTED]> wrote in message
> news:Q4HI6.118153$[EMAIL PROTECTED]...
> > > > Got one already. A BSCS from Rensselaer Polytechnic.
> > >
> > > You must be a minority, then. Nobody as stupid as you could graduate
> > > with a BSCS unless you are a member of some political "victim" group.
> > >
> > Atleast he finishes what he started, unlike some people.
>
> Hmmmm?
>
> What's all this?
>
> Is there an amusing story behind this comment?
>
I though it was amusing, Aaron calling some stupid for finishing college,
when Aaron never did...
------------------------------
From: Rick <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Justice Department LOVES Microsoft!
Date: Sun, 06 May 2001 18:15:20 -0400
Ayende Rahien wrote:
>
> "Rick" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Ayende Rahien wrote:
> > >
> > > "Rick" <[EMAIL PROTECTED]> wrote in message
> > > news:[EMAIL PROTECTED]...
> > > > Ayende Rahien wrote:
> > > > >
> > > > > "Roy Culley" <[EMAIL PROTECTED]> wrote in message
> > > > > news:[EMAIL PROTECTED]...
> > > > > > In article
> <dJZI6.6119$[EMAIL PROTECTED]>,
> > > > > > "Daniel Johnson" <[EMAIL PROTECTED]> writes:
> > > > >
> > > > > > > I *strongly* suggest that printing *must* be
> > > > > > > addressed as soon as possible. I honestly see
> > > > > > > very little movement on that front.
> > > > > >
> > > > > > I don't understand this printing problem that Unix supposedly
> > > > > > has. Unix apps normally produce postscript. Unix lpr uses filters
> to
> > > > > > see what the data type to be printed is. If you don't have a
> > > > > > postscript printer the filter will use ghostscript to convert it
> to
> > > > > > pcl or whatever. I now have a Lexmark z52 which has its own Linux
> > > > > > driver. To me it is totally transparent. Before I had an
> > > > > > HP850C. Again printing was totally transparent. Where is the
> problem?
> > > > >
> > > > > First, there is a news server that is dedicated to printing in
> Linux,
> > > that
> > > > > is a problem all in itself.
> > > > > Second, we aren't talking about end users' problems here, we are
> talking
> > > > > about support for developers.
> > > > > Let's say that we take two comparable products that does the exact
> same
> > > > > thing, one for windows, the other for linux.
> > > > > The windows one could have printing support in a matter of no time,
> > > because
> > > > > GDI abstract the output device from the developer, so you don't have
> to
> > > > > change the code at all to print to a printer or display on the
> screen.
> > > > > The only worry you have is page breaks.
> > > > >
> > > > > On Linux, however, you need to have seperate code that does it,
> which is
> > > > > much harder than adding few "if"s for page-breaks.
> > > >
> > > > Gee, I thought all I had to do was add a printer drive, and all a
> vendor
> > > > had to do was write a driver. Much like a printer driver is written
> for
> > > > any other OS.
> > >
> > > On user side, maybe, I really didn't deal with printing on linux on the
> user
> > > side.
> > > I'm talking on the *developer*'s side here.
> > > There is no abstraction on linux of the printer. In Windows, I don't
> *care*
> > > what printer it is, or if I print to printer or screen.
> > > The only thing that I need to change in my code to change displaying on
> the
> > > screen to printing is page breaks. On Linux, I would've to do all the
> > > formatting myself, and create a PostScript output, which is good for
> > > printers that support it, but what about those that don't? The other
> option
> > > is to support *myself*, every brand of printer in existance.
> > >
> >
> > I have an Epson 740. It doesn not support Postscript. Thats why I have
> > Ghostscript.
>
> Now, what does GhostScript *does*?
... and you are trying to pass yourself off as a software engineeer?
Ghostscript is a Postscript interpreter. You know, so you can use
Postscript output on non-Postscript printers.
--
Rick
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: How to hack with a crash, another Microsoft "feature"
Date: Sun, 6 May 2001 16:57:32 -0500
"Charles Lyttle" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Erik Funkenbusch wrote:
> >
> > "Charles Lyttle" <[EMAIL PROTECTED]> wrote in message
> > news:[EMAIL PROTECTED]...
> > > Erik Funkenbusch wrote:
> > > >
> > > > "Roy Culley" <[EMAIL PROTECTED]> wrote in message
> > > > news:[EMAIL PROTECTED]...
> > > > > In article <Ny7I6.22197$[EMAIL PROTECTED]>,
> > > > > "Erik Funkenbusch" <[EMAIL PROTECTED]> writes:
> > > > > >
> > > > > > I guess it depends on what you mean by "secure". If someone
doesn't
> > > > know
> > > > > > the decode algorithm, 4-bit encryption could be quite secure
> > > > >
> > > > > What crap. If you don't understand something don't make pathetic
> > > > > attempts to show that you do. ANY 4-bit encryption algorithm could
be
> > > > > cracked by brute force in less time than it took you to write such
> > > > > rubbish. The best known encryption algorithms are known and open
to
> > > > > peer review. If you invent a new encryption algorithm but won't
make
> > > > > it open to peer review then it just will not be accepted. Security
> > > > > through obscurity just doesn't cut it at any time.
> > > >
> > > > What's crap is your understanding.
> > > >
> > > > You can only brute force it if you know the decode algorithm. You
can
> > > > guess, and analyze and do lots of things, but it could be things
like
> > XORing
> > > > the data against a pets name, while rotating 3 bits and compressing
it
> > using
> > > > 10 different compression algorithms. The number of possible
> > combinations of
> > > > decode algorithms is limitless.
> > > >
> > > You aren't required to know the algorithm to crack encryption. You
don't
> > > care about the algorithm, you care about recovering the message. So
the
> > > attack has to create an algorithm that decodes the message. It doesn't
> > > matter if the algorithm is the "correct" algorithm or not. In fact,
> > > doing things such as you suggest often make a code easier to crack.
When
> > > you apply multiple compression algorithms, or multiple xor, the
attacker
> > > doesn't have to know how many times you compressed, he just has to
find
> > > one scheme to go from encrypted message to plain text.
> >
> > Ahh, but that's just it. Such a scheme typically needs to have a
"rosetta
> > stone" or some way to identify at least one character or word in the
data.
> > Suppose the encrypted data isn't plain text at all, but something that
is
> > based on a random character set chosen for the day it was encrypted?
You
> > need a point of reference, and without having that, you might as well
have
> > monkeys banging on keyboards.
> >
> No, you don't typically need a "rosetta stone". If you have such, then
> you can apply a "known plain text attack". But that isn't the only
> attack.
I said a "rosetta stone" *OR* some way to identify at least one character or
word in the data.
> > Typically, when trying to break encryption without knowing the
algorithm,
> > you either look for common algorithms, or you look for patterns that
match
> > known language patterns.
> You look for clues. Compression algorithms for example, will add
> information to the file that permits deduction of the compression
> scheme. So applying (by computer) tests for compression will very
> rapidly "back out" the compressions.
You're assuming that someone just ran a file through zip. That's not what
i'm talking about. There are many compression algorithms that you can apply
without creating file information. The output of the algorithm should
appear random if you don't know that it is compressed data.
> >If you disguise the language patterns by making
> > sure that even the same phrase doesn't create the same series of bytes,
then
> > you remove the ability to deduce a new algorithm.
> >
> But you can't do that with a 4 bit key. A 4 bit key means a cycle length
> of 16. So every 16 letters, or words, you can get repeats. These repeats
> will have spacing with a factor of 16. If 16 is the smallest factor,
> then I need only to test for key lenghts of 2 and 4, which totals out to
> 4 +16 = 20 keys.
No, a 4 bit key simply means that the key is 4 bits. The key may not use
factors at all, it might be the value that is XOR'd for example, or any of a
billion other ways those 4 bits might be used to encode the data. You're
making the critical error of assuming the use of a known algorithm, which is
exactly my point.
> Enigma was originally cracked without any knowledge of the algorithm,
> and it had a key length of 26^7 IIRC.
If that were the case, they wouldn't have needed to capture the enigma
machines.
------------------------------
From: Nigel Feltham <[EMAIL PROTECTED]>
Subject: Re: Article: AOL in cahoots with Compaq, HP to derail WinXP, .NET?
Date: Sun, 6 May 2001 23:31:19 +0100
What about an AOL-Linux package?
If they did this and ended up supplying a full Linux Distro on every
magazine and mailshot they currently put their CD on then MS won't stand a
chance - who will pay at least $50 for an MS os when every magazine has a
full rival os on the cover. The only way MS could compete would be to make
their OS free as well - when combined with an enforced company split the OS
company would soon be wiped off the map and the applications company would
be forced to code for Linux, FreeBSD, etc
------------------------------
From: Giuliano Colla <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,alt.linux,alt.destroy.microsoft
Subject: Re: The long slow slide to Microsoft.NOT
Date: Sun, 06 May 2001 22:21:10 GMT
Ayende Rahien wrote:
>
> "Giuliano Colla" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > JVercherIII wrote:
> > >
> > > I don't know too much about Delphi but I'll look into it. I'm still
> > > learning C++, Java, and Perl among other things in college. Pays to be
> > > diversified.
> > >
> > Well, in a nutshell, Delphi gives you the same WYSIWYG visual interface
> > of VB, but the code you write is Object Pascal instead of Basic, making
> > it robust, easy to maintain, and as fast as C++. Object Pascal is very
> > similar to C++, but, because of the strict type checking of Pascal, it's
> > easier to write robust code, and harder to make mistakes. In my opinion
> > it would be better to learn Delphi first and C++ after, because the
> > object handling is the same, but Delphi/Pascal is "cleaner", so you have
> > less Stroustrup whims and C heritage to cope with.
>
> Actually, I think that the reverse is true.
> I know a little of Pascal, and the pointers that Pascal have can stump you
> if you are transferring your Pascal knowledge to C/C++.
> Pointers aren't an easy concept in the first place, and C & C++ way of doing
> it are nearly identical to the way the machine does it, so it's easier to
> learn the C's way, and then learn Pascal, which limits your ability to use
> them.
> As a note, a good C++ programmer shouldn't have much use for raw pointers.
> The standard library provides for nearly everything that you need to use a
> pointer for, and it does it much better than the average programmer can hope
> to do. More safely, too.
> The big plus of C++ is that you don't have to pay for things that you don't
> need.
That's exactly my point. Pascal properly hides what shouldn't be
normally used by an application developer (i.e. dirty tricks with
pointers). You can do in Pascal anything you can do in C, but you must
state very explicitly, so that you're made aware of what you're doing.
Once you've learned, you may start playing with C++, if you feel like,
but your background will make you avoid all the trivial errors you can
do with C++, without the compiler telling you, and learning only at run
time.
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: Article: Want Media Player 8? Buy Windows XP
Date: Mon, 7 May 2001 01:09:46 +0200
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Ayende Rahien in alt.destroy.microsoft on Sun, 6 May 2001 13:33:48
> >"green" <[EMAIL PROTECTED]> wrote in message
> >news:9d38m6$r3r$[EMAIL PROTECTED]...
> >
> >> Probably wouldn't be hard to put a coffee maker on, but that may
violate
> >> some gpl on the howto for getting linux to make coffee.
> >
> >That is one thing that frighten me about the GPL.
> >There is already GPLed data, what happen when other things start to get
GPL?
>
> Then the world becomes a better place.
Really? Then why don't tell me how happy you would be when the credit card
companies would publish *your* credit history, because of GPL?
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: gnu.misc.discuss,comp.os.ms-windows.advocacy
Subject: Re: Richard Stallman what a tosser, and lies about free software
Date: Mon, 7 May 2001 01:20:25 +0200
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Les Mikesell in comp.os.linux.advocacy on Sun, 06 May 2001 03:10:38
> >"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
> >news:[EMAIL PROTECTED]...
> >
> >
> >> By "in theory" I meant to exclude such moronic ideas as those coming
> >> from you right now. This "here is code" is THEORY. PRACTICE is when
> >> you make a real program that real people use.
> >
> >So ask any software company if in practice they wait for all components
> >to be finished before writing any other components that use their
> >interfaces.
>
> This isn't necessary for my claim to be true. Are you saying that one
> component is produced and delivered before being tested with the others?
Yes.
> A rather sloppy development model, eh?
No.
> One might almost suspect that
> the ability to pull that off would be dependent on the various parts not
> being *derivative* of each other, wouldn't you say?
No.
The ability to pull that off would be dependent on the various parts
followring their respective API.
> For example, if a developer were to write a program that required a
> library which had an API published, but for which no library was ever
> coded. The developer could certainly offer the product for sale, but
> would anyone buy it? Hell, could he give it away?
If it does something useful, of course he could sell it.
He would just have to write the library himself, or wait for someone else to
do it, and then he is first to market.
> Perhaps, if there were people who knew of this API, and anticipated the
> library. But until that library appears, there really isn't anything
> but purely academic (scientific; knowledge not copyrightable) value in
> the program.
It's most certainly copyrightable.
> Now, say, some enterprising young man, with honest effort
> and ernest intent, decided to be a good productive member of society and
> writes a library. He'd like to make money on it; the program's
> existence provides him a market, but the program author (one of those
> GPL nut flakes, so insane that he not only sides with RMS, he's nutty
> enough to agree with T. Max that a library can be derivative of a
> program!) threatens to sue him in court if he doesn't pay royalties.
Not going to work. Neither is a deriative of each either, at best, they are
both based on the API. And the library much more so.
> Now just explain to me, without wigging out into any fantastical
> posturing based on rhetorical claims, should the guy who writes the
> library pay the program writer royalties, or should the program writer
> pay the library writer royalties? The API is 'public property', because
> APIs aren't copyrightable, so it doesn't even matter which of them, or
> some other person, designed it.
I don't see any reason why you can't copyright an API.
But even assuming that the API is public domain, then the program & the
library aren't.
So, in this case, the library author can demand royalties for distibuting
his library with the program.
Or even using his library, though that is a bit tricky to pull off.
The program author has the choice of not ship his program with the library.
He can do distribue his program as a stand-alone, and say, if you want to
use this program, you need to buy/get/warez XYZ library, or any other
library that implements the ZYX API.
The only way the program author can come with demands to the library's
author is if he contracted the library's author to create this library.
> If you were either party, in turn, and
> wanted to get paid for your efforts, but were threatened with suit by
> the other, what set of 'rules' can you provide that allow the correct
> result, regardless of your perspective, and the uncertainty of the
> claims?
See above.
If I were the library author, I would say, you want to use my library, give
me Y$ per copy, or Z$ total, or X% of the money you make (and make sure it's
not from the profit, this has a tendecy to be very low in such contracts).
If I were the program's author, I don't see *how* I could sue the library's
author. And I would've four choices, agree to the demands of the library
author, write my own implementation, contract someone to implement the
library for me.
Or wait for someone else to write an implementation, and hope that they
would be more reasonable.
What Every Computer Consultant Needs to Know:
1) In case of doubt, make it sound convincing.
2) Do not believe in miracles. Rely on them.
--Murphy's Computer Laws (Finagle's Rules)
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: Why is Microsoft opening more Windows source code?
Date: Sun, 6 May 2001 17:05:33 -0500
"Ray Chason" <[EMAIL PROTECTED]> wrote in
message news:[EMAIL PROTECTED]...
> "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote:
>
> >The problem is that the FSF and the Linux community in general keep using
> >the ambiguous term "free software", despite the fact that the FSF and the
> >Linux community know for a fact how confusing this term can be. All the
> >screams of "Free as in speech, not Free as in beer" are pointless.
>
> Yes, Stallman's insistence on "free software" instead of "open source
> software" strikes me as pedantic. Stallman can tell us why he has this
> preference till he's blue in the face; the two terms are still more or
> less the same in the real world.
Well, they're not the same really, and I understand that Stallman wants to
differentiate them. My beef is the use of the term "free", and it always
has been. They could have chosen lots of different terms, but then that
would be unambiguous, and would be a detriment to the GPL being snuck in the
back door of companies without their knowledge until it was too late.
> >Most of the people i've talked to about the GPL, many of whom were
planning
> >to release code under the GPL really had no idea what the full
implications
> >of it are. They simply think that it's no-cost or public domain
software.
>
> This is vague. Are you assuming that, if I release a given source file
> under the GPL, I cannot then use it elsewhere in a non-GPLd project?
> If I hold the copyright, then that isn't so. It only applies to other
> people's code that is GPL'd, i.e. not when I grant, but when I receive
> a license using the GPL.
No. As an example, I was talking to the author of an open source c++ class
framework on Source Forge, he had the framework listed as having the GPL
license. I asked him about this, and mentioned that this would mean that
the framework could only be used in other GPL projects (by anyone other than
himself), and not non-GPL'd projects. He was surprised and had no idea.
> I point you to a company with a reasonable business model that turns on
> this fact. Troll Tech licenses Qt as open-source software, under the
> licensee's choice of the GPL or something called the QPL (unique to Qt
> AFAIK), and also as a closed-source library that may be used in closed-
> source programs. The point is that if you're doing GPL'd code, you can
> have Qt for free, but if you're doing paid software, Troll Tech gets a
> license fee for it.
This is fine, so long as you own all of the software. But suppose Trolltech
had 1 line of GPL'd source code in their library. Suddenly, they no longer
have the choice to relicense it any other way.
> Alas, I don't know whether Troll Tech is making any money, but they've
> been around for a few years.
>
> Furthermore, again quoting the GPL:
>
> These requirements apply to the modified work as a whole. If
> identifiable sections of that work are not derived from the
> Program, and can be reasonably considered independent and separate
> works in themselves, then this License, and its terms, do not
> apply to those sections when you distribute them as separate works.
> But when you distribute the same sections as part of a whole which
> is a work based on the Program, the distribution of the whole must
> be on the terms of this License, whose permissions for other
> licensees extend to the entire whole, and thus to each and every
> part regardless of who wrote it.
>
> Thus, it is not the intent of this section to claim rights or
> contest your rights to work written entirely by you; rather, the
> intent is to exercise the right to control the distribution of
> derivative or collective works based on the Program.
However, the term "derived work" is not defined in the GPL.
> [EF]
> >> >Right after, an edict was issued that *NO* code that originated
outside
> >the
> >> >company could be used in any product, no matter what its license.
>
> [RC]
> >> This was a paranoid overreaction; plenty of free/open source software
> >> exists that is LGPL'd, or uses different licenses; code that is meant
> >> to promote a standard, such as libpng, probably should use the BSD
> >> license or something like it.
> >
> >It may be paranoid, and an overreaction, but the issue has never happened
> >again. That's all they care about.
>
> A perfect example of why "suit" and "PHB" are almost always
> pejorative.
Do you really expect everyone in the world to be an expert on license
issues?
------------------------------
From: [EMAIL PROTECTED] (Richard Thrippleton)
Subject: Re: The _one_ thing that pisses me off about Linux
Date: Sun, 6 May 2001 23:23:11 +0000
In article <9d4gm4$iug$[EMAIL PROTECTED]>, kosh wrote:
>Richard Thrippleton wrote:
>
>> First, let's get one thing straight; I'm not a wintroll. I don't
>> even run windows on any machine I own. I am a committed a Linux advocate
>> (and evangelist). But I am gonna rant (and there is foam coming out of my
>> mouth). What pisses me off about quite a significant amount of Linux
>> software is the ridiculous x86 bias, ridiculous to the point of ignorance
>> of the existance of other CPUs.
>> On many occasions, I have wasted much bandwidth and time
>> obtaining sourcecode to some killer app, only to find that it will only
>> compile on x86, and there is f***all warning in previous documentation.
>> There seems to be some kind of arrogant view that Linux is only run on
>> x86 (I use PPC). Some lamer has put in an ASM hack; that's fine, but would
>> it really be too much trouble to put in a C substitute, or at least some
>>>warning< next to that big hunk of megabytes large source code?? I
>> have to use a modem half the year and this isn't helping!
>> All I want is fair warning that they haven't taken into account CPUs
>> other than their own. So those of you who know (or are) developers,
>> _please_ bang some heads and let people know that there are people who do
>> not use x86s. Oh, and let them know that some people have their ints
>> ordered the _right_ way round :>
>>
>> Richard
>>
>
>The answer is probably simpler then what you think. Some of the programs I
>have written may not run on windows or on other cpu platforms. However I
>don't know if that is true because I don't have those other platforms to
>test it with.
I can obviously understand that something coded on Linux/x86 won't
go across to Windows, or maybe not even FreeBSD/x86 . But if it won't
compile on Linux/PPC, then that is likely because there is some ASM
optimisation; any fool knows that x86 code won't run on PPC/ARM/68k. I'm
not saying ASM is a bad thing. It sure increases speed, but having a C
alternative or a visible warning before download would be polite. Or it
could be endian type problems; fairly predictable if you're doing something
like loading data files in int or higher size chunks (data file made on
small-endian machine) or low level hacking of data structures like making a
char size pointer to manipulate an int. I'd expect any developer to at least
be aware that other CPUs have different orderings, and to be aware of the
implications of such. Besides, endian fixes are, I believe, quite trivial
(take a look at Nuq from www.quakeforge.net). By the way, this isn't the
whole story of CPU porting problems, but then I know sod-all about the
problems of different size integers from platform to platform.
Richard
------------------------------
** 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 by posting to comp.os.linux.advocacy.
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
******************************
