Linux-Advocacy Digest #119, Volume #35 Sun, 10 Jun 01 21:13:02 EDT
Contents:
Re: Justice Department LOVES Microsoft! ("Daniel Johnson")
Re: Justice Department LOVES Microsoft! ("Daniel Johnson")
Re: More micro$oft "customer service" ("Daniel Johnson")
----------------------------------------------------------------------------
From: "Daniel Johnson" <[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: Mon, 11 Jun 2001 00:41:38 GMT
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Daniel Johnson in comp.os.linux.advocacy on Sun, 10 Jun 2001
[snip]
> >> Actually, Sun doesn't use "churn" to monopolize Java, as MS does on
> >> their platform, so you are, in fact, mistaken.
> >
> >Sure they do. They keep changing the damn thing- especially
> >by adding new APIs that other Java vendors must
> >implement.
>
> What other Java vendors are required to implement Sun's APIs?
They all most implement what is in java.*. This is pretty
much everything Sun adds.
> I'm not
> familiar with those details. Likewise, adding APIs is not 'churn'.
> Changing APIs are, and Sun does not churn their APIs, changing them for
> no other reason than to deter competition. Microsoft *does* do this,
> this is a *fact* and we *know* they do this because they documented it.
> Doh!
MS is quite good about maintaining compatibility with
their old APIs, actually. The "churn" they are accused
of is *adding* lots of APIS, and of this they are quite
clearly guilty.
> >This is just exactly what MS does with Windows.
>
> Whoops! You're wrong again!
No, really. MS is just always adding stuff.
> >It means that Sun always has the latest and greatest
> >stuff *first*.
>
> It would, if they did this. They don't. Bwah-ha-ha-ha-ha! Kind of
> leaves you high and dry for an arguments, doesn't it? <*chuckle*>
They certainly do it. It's not exactly a secret; go look
at Sun's website. They document the stuff they add.
[snip]
> >> >You seem very sure that there is *no* overlap
> >> >between what a "profit seeking competitive firm"
> >> >might do and what a "criminal monopoly" might do.
> >>
> >> That is rather more astute of you to notice than I would have given you
> >> credit for, to be honest.
> >
> >You seem unable to justify this presumption, as well.
>
> Then how is it you figured it out without my having to explain it to
> you, just from reading my words? You don't seem to have a firm grip on
> what "justify" means.
Ah. So "justify" means "express" in MaxSpeak.
I meant by it "provide evidence or argument for", not
"express"
Is that clearer?
[snip]
> >> They would work, but only competitively. If MS were to attempt to
> >> employ *any* competitive strategy, at any time (except what we might
> >> consider to be 'by accident'), then it would allow competition, thus
> >> destroying their monopoly power.
> >
> >How so? Why would such actions 'allow competition'?
>
> Depends on the actions. Doh!
You seem sure there are no actions which do not
share this property. Is this simply a coincidence?
[snip]
> >I suspect that once again you are using MaxSpeak(tm). Let
> >me try to guess your meaning.
> >
> >I'm guessing that "anti-competitive" means "results in
> >monopoly"; "competitive" must mean "results in
> >no monopoly".
>
> Anti-competitive means "not competitive". It turns out, that it does,
> as you have discovered, result in monopoly, but that doesn't have
> anything to do with the definition of the word. That is just how
> economics works.
No, this does not work. You could do things that
are competitive and other things that are not. Your
conclusions don't follow using this definition,
no matter what "competitive" means to you.
[snip]
> >If I got it right, then I think it is now your
> >turn to explain why "anti-competitive"
> >strategies are bad, or "competitive" ones good.
>
> Competitive strategies are good because when multiple producers have to
> compete on an open market, the most efficient producers will sell more
> goods.
You haven't said what "competitive" means to you yet;
if it means "results in no monopoly", then it is not
clear why "competitive" actions do result in any
of these things. Raw incompetance would be
"competitive", for instance, and it's hard to see
how such a thing would produce any such
results.
> Those who build things efficiently (without excessive waste,
> with minimal cost, without inconveniencing the consumer) make more
> profit than those who don't, and those who don't eventually run out of
> capital and may 'go out of business', or simply learn to be 'more
> efficient', or change the nature of the efficiencies by finding a new
> market (a new prospective consumer base) for their product.
IMHO this sounds a lot like the software
industry today. :D
> Anti-competitive strategies are bad because they do not result in this.
What if they do result in something good? What
if they result in better products than "competitive"
strategies do.
> Thus, there is inefficiency, waste. Inevitably, this uses up resources
> faster than necessary, the resources run out, and society dissolves into
> barbarism.
That, I think, is quite a strange view of the ultimate
consuences of monopoly.
> Really. I know it sounds overly political, but what can I say? I'm a
> big fan of capitalist free markets!
No, I wouldn't say so. Not unless "capitalist" and "free"
have special meanings for you, too. You are a big fan
of government-directed markets.
[snip]
> >> So MFC is covered by a license similar to GPL, forcing anyone
developing
> >> an app relying on MFC components to be covered by a Microsoft license?
> >
> >No, it isn't. MFC is under a typical MS EULA.
>
> What is it?
Long and boring. It's not entertaining enough
for me to post a copy.
> And why do you think an EULA is 'required' to use *source
> code*?
Well, some license is- that stuff is under copyright.
> It all sounds like quite a scam, to me; making developers agree
> to special end-user licenses.
Er, the developer is the "end user" in this case. MS still
uses the same acronym though.
MS does not say what license you must use to
license your own software, so long as you do not
go redistributing the source of MFC.
> >APIs are typically really low-level and often painful;
> >a framework papers over this, providing convient
> >access.
>
> So when an Application Programming Interface is not complete or correct
> enough to actually be used to program applications, you need to pay
> Microsoft for an EULA just to program an application?
No, it's more that when you want to program in
a higher-level language than C++, you want a framework
that will map the API into that higher-level alnguage.
Win32 is fine if you are doing raw C.
> >Frameworks are language specific; they permit the
> >programer to take full advantage of the language's
> >features, and they take care of mapping this to
> >the things the API understands- which are usually
> >quite limited.
>
> So basically it is a term used to 'paper over' the fact that Microsoft's
> support for various languages is Windows-specific
What the heck does this mean?
> and lock in those who
> want to use those languages to Microsoft's API?
No, no, with a framework you are supposed
to use the framework, not the API directly. You
can in principle rewrite the framework to
another API, actually- MS used to provide
a Mac version of MFC, for instance.
Though the resutls of doing this are usually a
bit unsatisfactory. Still, it is less lock-in than
you get if you write to the API itself.
> About what I figured.
Aw, you just made that up. :D
[snip]
> >Lots of people use it. It was available very early, and
> >a lot of code still uses it. Also it's well integrated into
> >MS's IDE, which is certainly a plus- for beginners, a big
> >plus.
>
> No, I meant outside the monopoly; non-MS, non-Windows, whatever.
Oh. It's rarely used outside of Windows.
> >It's still lagging behind its competitors substantially,
> >however.
>
> What competitors?
Borland's Delphi, for one.
[snip]
------------------------------
From: "Daniel Johnson" <[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: Mon, 11 Jun 2001 00:41:42 GMT
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Daniel Johnson in comp.os.linux.advocacy on Sun, 10 Jun 2001
> [...]
> >No. The turing test is not about computational power
> >or turing-completeness. It is a (proposed) test
> >for intelligence- or at least apparent intelligence.
>
> You don't seem to understand Turing's theories. Perhaps you
> concentrated too much on the math, and never even considered what he was
> actually saying.
What he was saying was, in fact, rather
mathematical. Turing only dabbled in AI
a little.
Most of his work is on the theory of
computation.
> >There is no particular reason to believe that
> >this is in any way related to computational
> >power at all.
>
> Well, I know that. I'm glad you know that. If you talk to the vast
> majority of people in computer sciences, however, this is not the case.
I think you are mistaken. This question is quite
controversial. There is not such an overwhelming
consensus as you describe.
> They believe that there are a wealth of particular reasons to believe
> this is related to computational power and the capabilities of the
> program. From the origins of information processing theory to Deep
> Blue, there are a lot of reasons to understand Turing's theories
Deep Blue has nothing to do with it. The origins of
"information processing theory" might, depending
on what you mean by that.
> (even
> without worrying about the math) do relate computational power to the
> turing test, very specifically.
What theory of Turings do you have in mind? Please
explain the relationship.
I suspect you have misunderstood the Church-Turing
thesis. I ask because I want to see if that is true.
[snip]
> >The term "turing machine" is not part of the vernacular
> >at all; it is comp-sci jargon.
>
> Well, I am going to disagree with you there. I will admit it is not
> widely used vernacular, but there is no shortage of people who deal with
> comp-sci jargon but are not computer scientists.
Such people may misunderstand the jargon and
misuse it, but you should resist the temptation
to do that yourself.
> > You may have a special
> >magic Max meaning, but in the field the meaning of the
> >term "turing machine" is very precise, and I am not
> >going to go along with your efforts to redefine this one.
>
> As with any specialty, there is a tendency to expect one's jargon to
> have magical properties, such as the inability to be used as vernacular,
> and a mystical ability to think your use more 'precise', even when the
> term is not being used within the specialty!
It's not magic. It's a question of knowing what the words
mean. They can be explained.
> For the 'meaning' of Turing machine, I turn to Turing.
That's probably a good idea.
> And following
> his work, the work of many cognitive scientists who have variously
> supported or refuted the information processing theory of the mind.
One would hope that they would educate themselves
on what a Turing machine is before trying to incorporate
it in their own theories.
If they do not, then they would be well advised to
avoid applying any of Turing's results.
But I'm afraid sloppy scholarship is not unknown.
For instance, some philosophers treat the Church-Turing
thesis as if it were proved, which it is not.
One hopes that these cognitive scientists you
read do not make such mistakes.
> The
> question of whether our brains work like computers, or computers work
> like our brains, is still open, despite your "I am a programmer so I
> know all the answers" blinders.
Sure. I did not suggest otherwise. Though the
question is usually whether brains are Turing
machines, or something like them. This is not
quite the same, though admittedly it's close.
[snip]
> >It certainly does not. A turing complete formalism
> >can express any computation that a turing machine
> >can perform.
>
> You're just begging the question. Kind of like with 'anti-competitive'
> and 'monopoly', but without the vital distinction of being correct.
> Whether a turing machine can be built from a turing complete language is
> not a question which has an answer.
You are mistaken. It's good question, but Turing answered it
years ago- the answer is "yes". Definitely yes, it has
been proved.
It is one of Alan Turing's most important
results, actually, in several senses.
Any turing complete formalism can represent a turing machine
interpreter- a so-called "universal turning machine".
This is because the term "turing complete" means
"able to represent any computation a Turing machine
can"; and what Turing proved is that a Turing
machine can represent such an interpreter.
>. In contrast, whether
> anti-competitive actions lead to monopoly does have a very concrete
> answer.
I suspect it has a trivial answer, because
you have redefined the terms. But I still
find it slippery, and I suspect your definitions
may not be internally consistant this time.
> >That's it. That's exactly what it means. We simply
> >do not know whether this implies "can do AI";
> >we don't know if a turing machine can do AI.
>
> Know, we do not know if a turing machine is possible,
We can make a fair guess; Turing mahcines have
unlimited memory- and this would seem to pose
quite a problem practically.
Probably they are impossible to really
construct.
>and we do not know whether AI is possible.
Indeed.
> They are, however, related definitions,
No, they are not.
>and the
> turing test, and the turing machine that passes it,
We do not know that it is possible for a Turing
machine to pass it, of course, even if we could
build one.
> is commonly
> understood (outside your specialty, perhaps) to embody 'artificial
> intelligence'.
This is more controversial than you think. Have
you not read of the Chinese Room argument?
Some people believe that the Turing test
is a good operational definition of an
artificially intelligent system. Some do not.
> A turing machine is something that passes the turing test, and is
> indistinguishable thereby from 'intelligence'.
No, it isn't. A Turing machine is a formalism for
describing computation- kind of a Ur-programming
language.
> Perhaps you computer
> scientists have morphed the term 'turing machine' over the last few
> decades to mean any computer (or perhaps the vernacular has been
> morphed, but it doesn't really make any difference) but AFAIK,
You definition of "turing macine" is peculiar to yourself,
Max.
> a
> language can be turing complete, but a computer cannot correctly be
> called a turing machine, until the requisite turing test for AI has been
> passed.
Bizzare. So you are saying that because you use the
term "turing machine" to mean "machine that passes
the turing test", you think that the original definition
is now *wrong*, and we must invent a new term
for that, just to please you?
[snip]
> >I don't know if Turing even models computers, in the
> >modern sense of that term. He was trying to model
> >algorithms. It's not clear that he succeeded in this.
>
> Thus, it is not clear whether computers are 'turing machines',
It is clear that they are not, but they seem to be close.
> and it is
> at least slightly mistaken for you to use the term, regardless of
> whether you were taught to do so by your professors or not.
I do not accept your personal authority to redefine
this term in order to avoid being embarrased.
> >It is provable that finite-memory computers are
> >less computationally powerful than turing machines,
> >though.
>
> Then why did you call a computer with a finite amount of memory a turing
> machine?
Did I? Sloppy of me. They are finite state automata,
not Turing machines.
[snip]
> >Emitting a DB/2 database is not wholly computational;
> >it is partly I/O. The format *counts*; you can't just
> >emit an equivalent structure to your own tape.
>
> Senseless babbling, I suspect. Again, this metaphysical 'computer which
> does not include I/O' is just an imaginary thing. You should try to
> avoid basing your reasoning on imaginary things.
No, you can build algorithm-evaluators that have
no I/O. They aren't Turing machines because
they have finite memory, but disallowing I/O
is possible.
[snip]
> >> I'm afraid the fact that every major component in Office were purchased
> >> falsifies your argument without a backwards glance.
> >
> >Word was not, for one.
>
> I'm afraid you're mistaken. It was bought years before Word for
> Windows, it was a DOS application, but it was Word.
No way that was the same codebase as the Mac
application that current versions descend from.
[snip]
> >You cannot explain, however, why, in your opinion,
> >MS never ever completes in a Max-approved manner-
> >even when this would be effective.
>
> You have not explained how it would be effective. You have merely
> presumed that somehow a monopoly can make more money by competing on the
> merits than by controlling prices and excluding competition. Needless
> to say, this is a rather silly notion on your part.
Why is it silly? You insist that competition is *never*
a profitable option for a company like MS, and dismiss
all examples of MS competing on merit.
You should *justify* your claim that competition
is incompatble with MS's business model.
> >It seems like you feel tha the Max-approved forms of
> >competition are all inferior and ineffective.
>
> No, it is the Supreme Court-approved forms of competition (nominally,
> superior products, business acumen, or accident of history) which are
> ineffective, when monopolization occurs. Which is, again, why
> monopolization was outlawed.
Fine, even if you arrogate to yourself the privilege
of speaking for the Supreme Court, you *still*
should show why those methods are ineffective.
And you might tell me why you think they
should be used, if they are so pathetic.
> >I honestly have to wonder why you are so hot
> >to trot on those methods, if tey are so lousy.
>
> You falsify your own argument when you try to base it on such silly
> straw-man claims. These methods are far superior to monopolization,
> because they enable one to make profit while still "failing" to dominate
> the market. Monopolization, on the other hand, is the weak man's way to
> *avoid* having to compete.
I didn't know yout hought Microsoft weak. But *why*
are these methods superior to monopolization? They
seem to be tickets to failure, in your view. What good
does anything else do if you just go out of business,
anyway?
> Now, tell me, if you believe that business routinely try to 'capture the
> market' and 'restrain trade' of competitors and that this is just part
> of doing business, then WHY IS IT that Congress thought to pass the
> Sherman Act outlawing these things? I mean, if they are just some
> private thing that 'Max came up with'?
I'm rather cynical about that. The Sherman Act is
a wonderful stick to beat wealthy companies with,
if they don't do what you want.
> >If, say, the only thing consumers care about
> >is distribution, then perhaps it *is* right
> >and proper to compete on the basis of who
> >can get what bundled with what.
>
> Perhaps. Were you expecting me to disagree?
Yes, rather. You seem to think that bundling is bad.
I guess it's only bad if MS does it, though.
> You seem to be assuming
> that because a single producer or a single vendor *could* be efficient,
> then we are supposed to take for granted that it is. That's seems
> rather stupid, doesn't it?
Who said anything about a single vendor?
Though historically, vertical integration has been
successful. But that's not what's going on in
this business.
> And again, we are provided a good clue about how to handle the question,
> by the statutes and precedent that make up anti-trust law. If
> anti-trust is 'outlawing competing aggressively', as you seem to think,
> then why is it that not every anti-trust defendant is convicted?
Perhaps they have better lawyers than MS. :D
The Sherman Act is just vague, it doesn't outlaw
anything specific. This means the DoJ can twist it,
but it *also* means that your lawyers might do
the same.
It's not foreordained that the government will
always win with the Sherman Act.
[snip]
> >> Obviously, if he could not overcome the "disadvantage" of being unable
> >> to compete, then his company could not.
> >
> >I don't follow. The "disadvantage of being unable to complete"?
>
> You don't think being unable to compete is a disadvantage?
I do. But you *don't*, or so you've said- you
seem to feel that companies that do not
compete do better.
[snip]
> >> It doesn't help your case to so flagrantly mis-state the facts. The
> >> government has laws against monopolization, engaging in
> >> *anti-competitive* actions.
> >
> >Which, I observe, includes competing in markets
> >you do not already dominate.
>
> Bwah-ha-ha-ha-ha! You act as if "markets" were firm and concrete things
> that are simple to differentiate and persistent.
Well, the *law* talks that way. Of course it's
not true- that is one of the big things that's broken
about the Sherman Act.
> If you have a monopoly, you can dominate any market you can tie to the
> monopoly market. Thus, the crime of 'tying', which is really a
> violation of section 1 of the sherman act, which outlaws contracts in
> restraint of trade, not bundling per se.
You say that as if it were easy. How come MS
sometimes fails?
[snip]
> >> That was the journey. Doh!
> >
> >Well, part of it.
>
> You are making up a metaphoric journey to begin with.
Of course.
> You've provided
> no description of any other part, and you have not refuted my claim that
> it was, in fact, all the parts.
I did not think you wer emaking that claim
seriously. It's absurd on the face of it.
> My point, obviously, was that there
> was no 'journey', metaphoric or otherwise, and you are just making up
> random babbling again to try to move the discussion off the topic.
So MS did not manage to transition to Win32,
then? There's no such product as Windows 2000?
[snip]
> >You just willfully ignore the things MS does *right*
> >on a technical level- even when I point them out to
> >you specifically, and compare them to their competition.
>
> What competition? You honestly think that a two-bit company without any
> OS monopoly is going to magically be able to force MS to compete on the
> merits? Just because you can pull out some random features that you
> like is hardly convincing proof of competitive merit.
It's considerably better than what you have
to offer, Max.
> Every example you
> provide merely begs the question, "well, if they can be competitive
> then, why do they spend so much time and money trying to derail
> competition?"
I think that's been answered, you know. MS knows
that technical merit isn't everything, and they
are aggressive- they want to win by *all* means.
This includes, for instance, aggressive distribution
policies. And maybe worse thigns- they've been
accused of some *real* misdeeds.
But it *also* includes making better
products.
[snip]
> >> As if I cared.
> >
> >I do realize, Max, that you do not actually care
> >whether Windows is better or worse than this
> >or that other OS, on a technical level. Thus, my
> >technical comparisons are of no interest to you.
>
> You're finally catching on that far, at least.
Hey, progress! :D
> Someday maybe you will
> understand the truth. Windows is simply not capable of being better on
> a technical level unless it has to compete, so that the non-unanimous
> consensus of a free market can determine what "better" means when it
> comes to whatever the product is.
I say this has already happened, and the market's
verdict was that Windows was *much* better.
I present the technical details to show that this
is not absurd, as is sometimes claimed; Windows
is not just "crapware", it's got demonstrable advantages
and strengths.
> IOW, you are essentially imagining
> MS's technical value, because you haven't any feasible alternative to
> compare it to.
Sure I do. I have compared it in this thread to
MacOS and Unix.
> If ANY product were to actually compete with MS, then MS
> would pail in comparison; either too expensive, or too unreliable, or
> too unacceptable in some other ways.
If some other product were better than Windows, then
Windows would be worse than it, yes. But I say
that the products that do compete with it are not better.
> So the fancy features and acronyms
> you pretend make monopolization a good thing are just your own private
> delusions, and not technical comparisons of any interest to anybody.
They are of interest to people who believe
the quesiton of which OS is better is meaningful;
you seem to feel that it is not.
But that is not very consistant of you; you condemn
Windows as "monopoly crapware", as if it were
objectively bad. Your own argument would indicate
that it cannot be so, because nothing is good or bad
in your world, except by the verdict of a market
that is divided.
> >But I think others might read these postings,
> >possibly. :D
>
> If you know that, why are you smiling? Were I you, I would be rather
> embarrassed by that fact. Do you honestly think this exchange has made
> either you or Microsoft look any better?
Yes.
I think that the case that MS Windows are in
some *techncial* sense better should be made;
after all, very frequently the technical *faults*
of Windows are stressed, and this gives
an unbalanced view, to say the least. Pointing
out what they got right that others got wrong
does more for MS's case than just saying
"it is so as stable as Unix", which is what
some people do.
As for me, I confess I enjoy lecturing you
on these things. It's a ego trip. :D
------------------------------
From: "Daniel Johnson" <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy
Subject: Re: More micro$oft "customer service"
Date: Mon, 11 Jun 2001 00:42:57 GMT
"drsquare" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Sun, 10 Jun 2001 14:47:44 GMT, in comp.os.linux.advocacy,
> ("Daniel Johnson" <[EMAIL PROTECTED]>) wrote:
[snip]
> >HTML is just not the right tool to use if you
> >want that kind of control.
>
> What would you use then? A giant .gif?
How about PDF?
------------------------------
** 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
******************************
