Warning to newbies

[spinoza1111]

Beware of certain buzzwords including "sequence points" and "trap
representation".

They have no scientific content, and instead were developed to cover
up the inadequacies and very impossibility of "standardizing" a toxic
waste dump (the C programming language and its dialects).

Their use indicates intellectual fraud.

A "sequence point" is when a "standard" C compiler "must" evaluate.
The existence of the buzzword is a cover up of the fact that the
standards committees consisted of people more concerned with vendor
profits who had no remit to determine a standard semantics and a
rational evaluation order, because they were afraid of discommoding
vendors.

A "trap representation" is a pointer in some sort of theological state
of sin that points **** knows where. The "C standardization"
philosophy is that we should close our eyes in holy dread and weave a
circle 'round it thrice when in fact in calculating a pointer, an
intermediate value might not be a legal pointer. The simplest case is
the fact that you usually don't want to point at memory location 0.

C standardization is pseudo science and snake oil. Please don't get
taken in.

 

[jacob navia]

spinoza1111 a écrit :

Beware of certain buzzwords including "sequence points" and "trap
representation".

They have no scientific content, and instead were developed to cover
up the inadequacies and very impossibility of "standardizing" a toxic
waste dump (the C programming language and its dialects).

Idiot

 

[Nick Keighley]

Subject: Warning to newbies

Beware of certain buzzwords including "sequence points" and "trap
representation".

to those who aren't aware of it spinoza has some sort of axe to grind.
A troll in other words.

They have no scientific content, and instead were developed to cover
up the inadequacies and very impossibility of "standardizing" a toxic
waste dump (the C programming language and its dialects).

these terms are well defined by the C standard. If you interested in
their defininition then look them up in the standard. If you are
interested in their practical value then check out out past posts in
this newsgroup or ask!

[rougly speaking:
sequence point: a point in the source code where the compution must be
completed. In between sequence points there may be a choice as to the
order in which various sub-computions can be doen. This gives
implementors freedom to reorganise code foroptimisation purposes.

trap value: an illegal value. Reading such a value may terminate the
program. TVs are rare in integer formats but many floating point
formats support bit patterns that are not actaully floating point
numbers (infinities and NaN (not a number) values.
}

 

[Michael Foukarakis]

Beware of certain buzzwords including "sequence points" and "trap
representation".

They have no scientific content, and instead were developed to cover
up the inadequacies and very impossibility of "standardizing" a toxic
waste dump (the C programming language and its dialects).

Their use indicates intellectual fraud.

A "sequence point" is

A "trap representation" is

<snip>

Please refrain from using the aforementioned expressions, as you risk
being held liable to charges of crimes against public disorder,
irrational thought infringement, newsgroup trespassing, verbal
assault, code you have never written but fails to (inter-)operate
anyway, resistance against or obstruction of common sense, unlicensed
possession of buzzwords, brain abuse and molestation, keyboard misuse,
unlawful detention of newbies, possession of illegal arms,
contributing to delinquency of non-regulars, stalking, negligence and
other forms of not knowing what the **** you're talking about,
alternate reality definitions of nonexistent expressions which invoke
undefined cosmic behaviour under the C0x99 Standard (SI, not ISO),
first degree brain cell murder and crimes against humanity as defined
by the Hague Statute of the ULD and outlined in the Proclamation of
Inherent Powers. Do note that all affected parties are hereby
considered being served an implied notice of aforementioned
activities, and all pertaining actions shall be implemented in the
pursuance of the related objectives, most notably Ownage of the
Defendant (aka. spinoza1111). You shall also receive formal written
announcement communicating scheduling information about your
involvement in said offenses. The original notice has already been
filed (against all parties involved) with Mr. Syndrome, Internal
Intern and expert on the field of Hypothetical Malpractice of the ULD,
A.I. Chains, Congressional Liaison and expert on Hogwartsian Law,
Vandal and Lurker Profiling, as well as Department Tail, Dr.
Happytimes, Moral Bankruptcy Law Specialist. Concluding, I remark you
have the right to abandon your keyboard immediately and the obligation
to remain silent, as anything you say slash type slash unsuccessfully
try to communicate can and will be used against you in accordance to
Usenet Law.

 

[Colonel Harlan Sanders]

spinoza1111 a écrit :

Idiot

Of course you're correct, but:

Colonel Harlan Sanders a écrit :
[snip off topic polemic]
Look, here is a C group.
You do not like somebody?
Use private email, blog, whatever.
You do not like spinoza111?
DO NOT ANSWER.
Let's discuss about C ok?

Try taking your own advice, rather than handing it our so freely and,
may I say, obnoxiously.

Nilges was clearly laying troll bait, carefully crafted to prod his
usual nemeses into engaging with him.

 

[Seebs]

That isn't what a troll is. A troll is someone who posts deliberately
provocative material, the objective being to incite a hostile reaction
for the heck of it. A true troll has no axe to grind, just a newsgroup
to pester.

As someone who spent years reading alt.religion.kibology, I'd like to
point out that, much like "hacker", the term has more than one sense.
Trolling is a kind of fishing; instead of throwing a baited hook where
you think the fish are, you move your boat around with a baited hook
in the water behind you, and some fish go for it because it's moving.

Generically, "trolling" can be used for any activity designed to provoke
*any* kind of responses, as long as the goal is the responses in and of
themselves. Asking a question because you want to know the answer isn't
trolling; asking a question because it would amuse you if people answered
it probably is.

It is worth noting that, in some cases, trolling is directed not at a
newsgroup, but at a specific person, and that a skilled troll can be an
asset to a newsgroup or forum. On some of the web-based bulletin boards
I hang around on, I've seen trolls do a very good job of dealing with
obnoxious nuisance users, by posting things that everyone else would ignore
but which would tie the nuisance users up for hours -- this having the
convenient effect of, say, keeping the nuisances from harassing people who
were actually hurt or offended by their behavior.

That works better on a forum where threads you aren't interested in don't
have to be "skipped", you just don't navigate to them.

On the other hand, many users can enjoy a successful and interesting troll
played for comedy value. Back in the day, there was a long-running thread
about whether a given number was prime; all I recall is that it was about
twelve digits, the last of which was an even number. The various pseudo-math
offered to "prove" that this number was prime was funny. (I still have fond
memories of discovering that, yes, the Internet contains people who can be
convinced that ATMs print money rather than having a supply of pre-printed
money. The best part was someone who worked at a bank, and testified that
his job included putting fresh rolls of paper in the ATM. When someone
said those were for receipts, a third party jumped in and said "Don't be
ridiculous, those are preprinted.")

-s

 

[spinoza1111]

spinoza1111a écrit :

Idiot

Of course you're correct, but:

Colonel Harlan Sanders a écrit :
[snip off topic polemic]
Look, here is a C group.
You do not like somebody?
Use private email, blog, whatever.
You do not like spinoza111?
DO NOT ANSWER.
Let's discuss about C  ok?

Try  taking your own advice, rather than handing it our so freely and,
may I say, obnoxiously.

Nilges was clearly laying troll bait,  carefully crafted to prod  his
usual nemeses into engaging with him.

No, I'm discussing C. I suggest you do so as well.

 

[spinoza1111]

As someone who spent years reading alt.religion.kibology, I'd like to
point out that, much like "hacker", the term has more than one sense.
Trolling is a kind of fishing; instead of throwing a baited hook where
you think the fish are, you move your boat around with a baited hook
in the water behind you, and some fish go for it because it's moving.

Generically, "trolling" can be used for any activity designed to provoke
*any* kind of responses, as long as the goal is the responses in and of
themselves.  Asking a question because you want to know the answer isn't
trolling; asking a question because it would amuse you if people answered
it probably is.

It is worth noting that, in some cases, trolling is directed not at a
newsgroup, but at a specific person, and that a skilled troll can be an
asset to a newsgroup or forum.  On some of the web-based bulletin boards
I hang around on, I've seen trolls do a very good job of dealing with
obnoxious nuisance users, by posting things that everyone else would ignore
but which would tie the nuisance users up for hours -- this having the
convenient effect of, say, keeping the nuisances from harassing people who
were actually hurt or offended by their behavior.

That works better on a forum where threads you aren't interested in don't
have to be "skipped", you just don't navigate to them.

On the other hand, many users can enjoy a successful and interesting troll
played for comedy value.  Back in the day, there was a long-running thread
about whether a given number was prime; all I recall is that it was about
twelve digits, the last of which was an even number.  The various pseudo-math
offered to "prove" that this number was prime was funny.  (I still have fond
memories of discovering that, yes, the Internet contains people who can be
convinced that ATMs print money rather than having a supply of pre-printed
money.  The best part was someone who worked at a bank, and testified that
his job included putting fresh rolls of paper in the ATM.  When someone
said those were for receipts, a third party jumped in and said "Don't be
ridiculous, those are preprinted.")

-s

Why is it, Peter, that you can only write coherently when you're
mocking other people?

 

[Eric Sosman]

Beware of certain buzzwords [...]

Please refrain from using the aforementioned expressions, as you risk
being held liable to charges of crimes against public disorder,
irrational thought infringement, newsgroup trespassing,[...]

"I swear to God I will see you in court."

 

[spinoza1111]

Subject: Warning to newbies



to those who aren't aware of it spinoza has some sort of axe to grind.
A troll in other words.

I'm not a troll. I'm a software developer with most of the Master's
degree in CS complete with a straight A average, thirty years of
experience, who's assisted John Nash and Jon "The Fate of the Earth"
Schell with C and the Mac, who's published on CS since 1976. However,
I also don't lie and I make dishonest and silly people uncomfortable.

these terms are well defined by the C standard. If you interested in

The C standard is the problem, because the C "standard" is bogus
science.

their defininition then look them up in the standard. If you are
interested in their practical value then check out out past posts in
this newsgroup or ask!

[rougly speaking:
sequence point: a point in the source code where the compution must be
"Compution"?

completed. In between sequence points there may be a choice as to the
order in which various sub-computions can be doen. This gives

"Computions". Once is a typo. Two is a subliterate trying to tell me
something.

implementors freedom to reorganise code foroptimisation purposes.

This is absurd. Had the members of the C standards board been
qualified they would have realized that it is not the language's job
to "help the optimizer". We know how to optimize WITHOUT changing the
order of computations in source code so as to get different results at
different times, and no other major language was designed or
redesigned "for optimization".

For example, the only way to "optimize" a+b+c correctly is to use the
commutative law. Whereas the members of the C standards team or group
(one of whom, Peter Seebach, had never taken a computer science class
and paid his way onto the group to advance his career) actually
believed that the language had to allow changes to evaluation order to
be optimized. This is the reverse of the truth: languages with
stricter rules are EASIER to optimize as long as you optimize in the
only ethical way possible, eg., preserving mathematical correctness.

trap value: an illegal value. Reading such a value may terminate the

To paraphrase Dijkstra: The problems of language standardization,
which is nothing more than language design, are much too difficult for
people who think in vague and corporate ways, compounded with sloppy
English.

Do you even know what "reading" a "trap value" might be? If I can
assign a pointer to void it's been read, and I can. The dishonesty of
the C standard is that it legislates against bad practice without
empowering compilers to detect it at run time, because the C standard
was developed SOLELY to enable vendors without any effort to label
existing compilers "standard".

Reading ANY value, not just values in this poorly defined subset, may
terminate the program, therefore any value is a trap value: the
concept is NOT part of computer science, it is voodoo hoodoo developed
by psychology majors actually proud that they've never taken a CS
class.

program. TVs are rare in integer formats but many floating point
formats support bit patterns that are not actaully floating point
numbers (infinities and NaN (not a number) values.

NAN is a floating point number whose use causes an interrupt (the
clown who called them "trap values" was probably some incompetent
geezer that vaguely remembered when interrupts were called traps). The
use of infinity and NAN doesn't produce "undefined" results in
sensible environments at all: if NAN occurs in an expression, the
expression is NAN, not undefined, and the same is normally the case
for infinity.

It is undefined in your mind:
Hand-waving and voodoo in the service of money is not science.
What you call "undefined" is, we find,
The name and only the name of your stupidity, greed and ignorance.
You learned in corporations intellectual dishonesty
And that form of male bonding called normalized deviance
Which is also evident in your bullying and ungracious uncharity
Towards strangers which was the sin of Sodom by chance.
You sat on your ass and you compromised,
And the evidence is in words which have no content,
Schildt took one look, and sighed,
These clowns are doing nothing important.
The only way to standardize C was to be formal and not undefined on
its semantics,
Which could have been elegantly defined in C.
But thugs in the room came from the shadows,
And said you have for this no authority.
Thou shalt not use your brains, atrophied as they were, and are:
Instead thou shalt do as money decrees if in this business, you would
go far!

 

[spinoza1111]

That isn't what a troll is. A troll is someone who posts deliberately
provocative material, the objective being to incite a hostile reaction
for the heck of it. A true troll has no axe to grind, just a newsgroup
to pester.

There is such a thing as a clever troll, but these are rarely seen in
comp.lang.c nowadays.

Heathfield is right.

 

[spinoza1111]

spinoza1111a crit :



Idiot

No, Jacob, I'm not an idiot. And if you'd not be bullied, don't bully
in turn. I've programmed in several languages successfully, and I
realized in 1991 that C was overrated because it allows smart people
to make stupid mistakes in service of providing an old-fashioned form
of computing "power" that is for the most part extremely marginal
today.

This is the Walter Mitty fantasy that the programmer is somehow in
reality assisting, if not second guessing, the "real man" hardware
engineer by squeezing cycles using a language which violates rules
made for lesser men. It allows programmers to avoid thinking in the
form of better problem definition and algorithm research.

It supports the nonsensical hacker "ethic" mythos that fat,
unimaginative and uncreative corporate shitheads are in reality
creative artists when they create software that creates silly problems
because of the shortcuts those slobs have taken.

**** you, Monsieur. I'll certainly be less interested in defending you
against the thugs in this newsgroup since you have proven you're a
thug, who like Heathfield has prostituted himself in order to
commercially promote a product.

 

[spinoza1111]

As someone who spent years reading alt.religion.kibology, I'd like to
point out that, much like "hacker", the term has more than one sense.
Trolling is a kind of fishing; instead of throwing a baited hook where
you think the fish are, you move your boat around with a baited hook
in the water behind you, and some fish go for it because it's moving.

Generically, "trolling" can be used for any activity designed to provoke
*any* kind of responses, as long as the goal is the responses in and of
themselves.  Asking a question because you want to know the answer isn't
trolling; asking a question because it would amuse you if people answered
it probably is.

It is worth noting that, in some cases, trolling is directed not at a
newsgroup, but at a specific person, and that a skilled troll can be an
asset to a newsgroup or forum.  On some of the web-based bulletin boards
I hang around on, I've seen trolls do a very good job of dealing with
obnoxious nuisance users, by posting things that everyone else would ignore
but which would tie the nuisance users up for hours -- this having the
convenient effect of, say, keeping the nuisances from harassing people who
were actually hurt or offended by their behavior.

That works better on a forum where threads you aren't interested in don't
have to be "skipped", you just don't navigate to them.

On the other hand, many users can enjoy a successful and interesting troll
played for comedy value.  Back in the day, there was a long-running thread
about whether a given number was prime; all I recall is that it was about
twelve digits, the last of which was an even number.  The various pseudo-math
offered to "prove" that this number was prime was funny.  (I still have fond
memories of discovering that, yes, the Internet contains people who can be
convinced that ATMs print money rather than having a supply of pre-printed
money.  The best part was someone who worked at a bank, and testified that
his job included putting fresh rolls of paper in the ATM.  When someone
said those were for receipts, a third party jumped in and said "Don't be
ridiculous, those are preprinted.")

-s

You've just self-indulgently wasted our time with a post that
contributes NOTHING to the discussion apart from an old staple of
break rooms in corporations: the foibles of other people, as opposed
to the implied wisdom of the narrator.

People sit around and tell these stories when they in fact have no
autonomy. They are like racist jokes, since the purpose of telling
them is to imply, without any intellectual effort, that the speaker
knows everything that's worth knowing.

 

[Mark]

No, Jacob, I'm not an idiot. And if you'd not be bullied, don't bully
in turn. I've programmed in several languages successfully, and I
realized in 1991 that C was overrated because it allows smart people
to make stupid mistakes in service of providing an old-fashioned form
of computing "power" that is for the most part extremely marginal
today.

<snip>

Edward,

If, as it sounds, you gave up on C in 1991, why hang out in comp.lang.c?
The MS enthusiast who niggles at Mac users in Mac forums will be viewed
as a troll. The same is true of many atheists in religious groups.
Why, as someone who doesn't rate C, go to a C group?

It doesn't make much sense...unless you're trolling.

 

[Michael Foukarakis]

I'm not a troll. I'm a software developer with most of the Master's
degree in CS complete with a straight A average, thirty years of
experience, who's assisted John Nash and Jon "The Fate of the Earth"
Schell with C and the Mac, who's published on CS since 1976. However,
I also don't lie and I make dishonest and silly people uncomfortable.

None of those are mutually exclusive with the troll status you hold.

The C standard is the problem, because the C "standard" is bogus
science.

The C standard is not science. It does not claim to be science, let
alone "bogus" science. You are therefore talking about things that do
not exist. In that context, I believe the minotaurs should abandon C
for VB, and then you may focus on something constructive.

"Computions". Once is a typo. Two is a subliterate trying to tell me
something.

You are focusing on all things irrelevant because you have nothing of
substance to say. The true way of a troll.

This is absurd. Had the members of the C standards board been
qualified they would have realized that it is not the language's job
to "help the optimizer".

....and that is your opinion.

We know how to optimize WITHOUT changing the
order of computations in source code so as to get different results at
different times, and no other major language was designed or
redesigned "for optimization".

The only reason you know how to do that is because the language allows
you to.

For example, the only way to "optimize" a+b+c correctly is to use the
commutative law.

Incorrect. This is not a math class, anyway.

Whereas the members of the C standards team or group
(one of whom, Peter Seebach, had never taken a computer science class
and paid his way onto the group to advance his career) actually
believed that the language had to allow changes to evaluation order to
be optimized. This is the reverse of the truth: languages with
stricter rules are EASIER to optimize as long as you optimize in the
only ethical way possible, eg., preserving mathematical correctness.

Wrong. The strictest language, one that demands operations will appear
in machine code in the order specified in source code, cannot be
optimized in the limited context of the optimizer you are concerning
yourself with. Perhaps you need to do some studying first, because
being published since '76 hasn't helped you much.

To paraphrase Dijkstra:

Paraphrase all you want, it's still bs.

Do you even know what "reading" a "trap value" might be? If I can
assign a pointer to void it's been read,

That's a write operation, not a read.

and I can.

With help from the minotaurs, sure.

Reading ANY value, not just values in this poorly defined subset, may
terminate the program, therefore any value is a trap value: the
concept is NOT part of computer science, it is voodoo hoodoo developed
by psychology majors actually proud that they've never taken a CS
class.


NAN is a floating point number
"value"

whose use causes an interrupt

No, its use AND production causes an exception.

The use of infinity and NAN doesn't produce "undefined" results in
sensible environments at all: if NAN occurs in an expression, the
expression is NAN, not undefined, and the same is normally the case
for infinity.

The C standard adopts IEC 60559 definitions and conventions for
evaluation of mathematical expressions. Does it state the result, when
one or both operands are NaN, is undefined? And what is it with
"sensible environments" people seem to be invoking all the time? There
aren't any sensible environments, they're all figments of our
imagination!

 

[Seebs]

[...]

The C standard is not science. It does not claim to be science, let
alone "bogus" science. You are therefore talking about things that do
not exist. In that context, I believe the minotaurs should abandon C
for VB, and then you may focus on something constructive.

Indeed, I would never have thought of it as being "science".

The only reason you know how to do that is because the language allows
you to.

Perhaps more importantly, there are many cases in which better optimizations
are possible if you allow the order of some operations to vary, because you
know that you don't care about the difference -- but it may not be possible
for the compiler to know that you don't care.

Incorrect. This is not a math class, anyway.

Indeed. In particular, it is worth noticing that there are cases where
obvious mathematical properties do not apply to C.

((a * b) / c) != (a * (b / c))

is true for some a, b, and c.

Ooh, I like that one. Did you know, I also paid for my driver's license?
Clearly, a dishonest attempt to further my personal travel options.

That's a write operation, not a read.

Also, there's no such thing as a "trap value". There is such a thing as
a "trap representation". "value" and "representation" are not interchangeable
concepts!

With help from the minotaurs, sure.

Yes. I particularly like "assign a pointer to void", because I don't think
it's possible. (It may be possible to cast it to void, but a cast is not
an assignment. So far as I can tell, assignment can be done only if
you have an lvalue of the type, and you can't have a void lvalue.)

No, its use AND production causes an exception.

.... Which may well be ignored, depending on context.

The C standard adopts IEC 60559 definitions and conventions for
evaluation of mathematical expressions. Does it state the result, when
one or both operands are NaN, is undefined? And what is it with
"sensible environments" people seem to be invoking all the time? There
aren't any sensible environments, they're all figments of our
imagination!

Yeah. And actually, I don't think that, assuming IEEE arithmetic, either
NaN or either infinity is a trap representation. Trap representations are
pretty rare, and many systems don't have any.

-s

 

[Eric Sosman]

None of those are mutually exclusive with the troll status you hold.

Whatever his failings as a computer scientist (thirty years
to not quite finish a master's degree -- maybe in another thirty
he'll not quite get a clue), he's a skilled and successful troll.
Observe that he's trolled *you*, and ponder what that means.

 

[Ike Naar]

To paraphrase Dijkstra: The problems of language standardization,
which is nothing more than language design, are much too difficult for
people who think in vague and corporate ways, compounded with sloppy
English.

That is a very liberal interpretation of what Dijkstra said.
For the record, here's the original quote:
"The problems of business administration in general and data base
management in particular are much too difficult for people that
think in IBMerese, compounded with sloppy English."
(EWD498, "How do we tell truths that might hurt?", June 1975)

 

[Keith Thompson]

I'm not a troll.

[...]

None of those are mutually exclusive with the troll status you hold.

[...]

Are you trying to convince "spinoza1111" that he's a troll? Do you
expect to be successful?

Are you trying to convince the rest of us that "spinoza1111" is
a troll? Do you think that's necessary?

 

[spinoza1111]

[...]
The C standard is not science. It does not claim to be science, let
alone "bogus" science. You are therefore talking about things that do
not exist. In that context, I believe the minotaurs should abandon C
for VB, and then you may focus on something constructive.

Indeed, I would never have thought of it as being "science".

What you do is not science, however there still is a computer science.

Perhaps more importantly, there are many cases in which better optimizations
are possible if you allow the order of some operations to vary, because you
know that you don't care about the difference -- but it may not be possible
for the compiler to know that you don't care.

"So anthropomorphic thinking is no good in the sense that it does not
help. But is it also bad? Yes, it is, because even if we can point to
some analogy between Man and Thing, the analogy is always negligible
in comparison to the differences, and as soon as we allow ourselves to
be seduced by the analogy to describe the Thing in anthropomorphic
terminology, we immediately lose our control over which human
connotations we drag into the picture. And as most of those are
totally inadequate, the anthropomorphism becomes more misleading than
helpful." - Dijkstra

Compilers don't "know" jack. Instead, they should optimize while
preserving the semantics of what you've coded using mathematical and
computer-scientific realities. That the "standard" "allows" you to do
something has nothing whatsoever to do with this. Languages with
sensible and predefined execution order are EASIER to optimize than C.

Indeed.  In particular, it is worth noticing that there are cases where
obvious mathematical properties do not apply to C.

((a * b) / c) != (a * (b / c))

is true for some a, b, and c.

We know what these are. The fact that floating point numbers are of
limited precision is a scientific and mathematical fact that can be
anticipated. It does NOT license the myth that programmers are in a
different business.

Ooh, I like that one.  Did you know, I also paid for my driver's license?
Clearly, a dishonest attempt to further my personal travel options.

Well, did you pay the examiner to pass you after you failed to
parallel park? That is a better analogy, since you either contributed
or watched.

If you contributed, your contributions were worthless because you
don't have the applicable education.

If you did not contribute, then you unduly claimed authority wrt to
Schildt.

Also, there's no such thing as a "trap value".  There is such a thing as
a "trap representation".  "value" and "representation" are not interchangeable
concepts!

This is three card monte: there is when you want it, there isn't when
you don't. There are five million google hits for "trap
representation" and they are all or mostly about the C standard.

This is the corporate game of an ignorance that is in part feigned but
believable when feigned because of the sea of ignorance on which the
feigning ship sails.

Yes.  I particularly like "assign a pointer to void", because I don't think
it's possible.  (It may be possible to cast it to void, but a cast is not
an assignment.  So far as I can tell, assignment can be done only if
you have an lvalue of the type, and you can't have a void lvalue.)


... Which may well be ignored, depending on context.  


Yeah.  And actually, I don't think that, assuming IEEE arithmetic, either
NaN or either infinity is a trap representation.  Trap representations are
pretty rare, and many systems don't have any.

Because the standard is junk science, words can conveniently mean
whatever you like.