I am new to lisp and therefore I have a cuple of (stupid) questions:
why is it that after I set the value of a variable using

(setf a (make-array '(2 2) :initial-contents '((.1 .2) (.1 "abc"))))

I don't get nil when I do

(typep a '(array float)) ?

How should I set a to make the previous test return nil?

(upgraded-array-element-type 'float)

return there?

Regards,
--
Nils Goesche
"Don't ask for whom the <CTRL-G> tolls."

[46]> (upgraded-array-element-type 'float)
T

(make-array '(2 2) :element-type 'float)

you would actually get an array that can hold any type.  TYPEP's result
reflects this -- it tells you that A is the type of array you would get if
you asked for it to be specialized for FLOAT.

Since FLOAT actually represents the union of multiple types (SINGLE-FLOAT,
DOUBLE-FLOAT, etc.), it's pretty uncommon to have an array type specialized
for it.  You're more likely to get the result you want if you do something
like:

(typep a '(array single-float))

(upgraded-array-element-type 'single-float)

returns T as well in my copy of CLISP.  However,

(upgraded-array-element-type 'character)

is CHARACTER, for instance, so

(typep a '(array character))

will return NIL.

Regards,
--
Nils Goesche
"Don't ask for whom the <CTRL-G> tolls."

(setf myarray (make-array '(5) :element-type 'single-float))

and then I try to set the value of one element of myarray to a
string, like:

(setf (aref myarray 3) "hello")

an error will be triggered only if string is not a subtype of

(upgraded-array-element-type 'single-float)

which is in turn implementation-dependent. Is that right?

  The type you want is one of the float types.  The superclass float is not
  useful for specialization.  One of the types short-float, single-float,
  double-float, or long-float is what you want.  You must specify it with
  :element-type if you want the array to be specialized for that type.  (This
  is pretty clear from the specification.  Please do not try to guess how
  things work when there is a specification available.)  An array does not
  become specialized just because you stuffed it with objects of a particular
  type at one point -- the Common Lisp runtime system has no way of determining
  what you wuold like to do with it and whether specializing it without your
  request would break something.  Neither does it know that a specialized array
  you have just created will be the only binding of the variable -- it could
  figure it out, but it is easier for you to tell it if you know, anyway.

  Note that if you try to mix a floating point number and a string in an array
  that you have specialized for some floating-point type, you have lied to the
  compiler.  How and when it will exact its revenge is not specified.  Again,
  see the specification.  Even if you can lie to some compilers all of the time
  and all compilers some of the time, you cannot lie to all compilers all of
  the time.

  Note that the variable *read-default-float-format* holds the type that the
  Common Lisp reader returns if you use an unadorned floating point number.

As with declaring the types of variables, specifying :ELEMENT-TYPE
indicates a promise that you will never store any other type in the
location.  It is *not* a demand that the implementation verify that you're
keeping your promise, although some implementations will perform such
checks if you're running in "safe" mode rather than optimizing for speed.

Erik,

  I think you need to read the original question again.  The original
poster was not expecting Common Lisp to infer the type from the the
contents, but was rather puzzled that the TYPEP test for a more specific
array type specifier did not return NIL.

  My guess is that in the implementation in question, the types
(array float) and (array T) are the same (in other words, there is
no specialized (array float).  Since they are the same, the typep
test returns T.

  Look, e-mail works just fine if you have to go personal.

| I think you need to read the original question again.

  I do not.  Please let the original poster speak for himself.  Your guesses
  and inferences about what somebody else intended are ipso facto _false_.  Do
  not even bother to post anything like that.

| The original poster was not expecting Common Lisp to infer the type from the
| the contents, but was rather puzzled that the TYPEP test for a more specific
| array type specifier did not return NIL.

  Are you quite sure you understand what a user expects when he asks for a non-
  existing specialized array type after stuffing almost only "floats" into an
  array?  A natural interpretation of his question would be "Does this array
  only contain floats?", and that is _not_ how typep works.  Do you understand?
  In order to take the original poster on a journey of discovery away from his
  expectations, I thought he needed to be told of how the things he had _asked_
  for works, so he would realize that he should not go further down that path.

| My guess is that in the implementation in question, the types (array float)
| and (array T) are the same (in other words, there is no specialized (array
| float).  Since they are the same, the typep test returns T.

  Look, a typep test for (array <non-t>) will _never_ return true if the array
  has been created with :element-type t, which is the default.  His test will
  _always_ fail, simply because he is confused about what constitutes a type
  test and how an array is specialized when it is created.

  I really don't have time for this Mickey-Mouse business.  Read the fucking
  standard.  Just figure it out.  If you have to "correct" me in public, at
  least be _more_ correct than I am.  Otherwise, I have to fight the _wrong_
  correction by quoting from the standard or wasting huge amounts of time
  dealing with people as clueless as in the call-by-value case.  (People pay
  lots of money and spend years of their lives studying in order to raise the
  general level of discourse among practitioners of a discipline, but then we
  get scores of people who believe that not having a clue at all is acceptable
  and that people should give them a computer science education for free?)  You
  can spend the time researching things yourself instead of requiring me to do
  it to correct a correction that isn't.  Goddamnit, I try to help people with
  what I see as the underlying confusion, at least pay some fucking attention
  before you imply that I was missing the point.

  This forum is _not_ rewarding if you already have a clue.  What a goddamn
  waste it has become to try to help people who already think they know the
  answer, no matter how wrong it is.  If you are not prepared for all kinds of
  corrections, even to what you believe strongly to be true, when you are stuck
  (a strong indication that you have missed something), you should _not_ waste
  people's time by asking for help.  There have been more of these people on
  the Net recently, people who expect a simple answer to what they think is a
  simple question and who have no desire at all to _learn_ anything, who refuse
  to listen, who only want someone to debug their code for them, and I do not
  care much for them.

  I really thought the orginal poster wanted to learn something, but if I am
  mistaken in this, I profoundly regret wasting the time to help him.  All in
  all, it looks like people who come to the Net these days are not worth
  helping, because they will never assume the role that the older generation
  assumed after they learned from others -- helping the next generation.  I
  answer people's questions for free on the Net because I think the world is a
  better place if people know their stuff and knowledge is shared among those
  who want to learn, but if the people who read the answers do not agree with
  this and only abuse my good-will to steal my time, it is not worth responding
  to anyone, anymore.

Oh wait, I figured it out. It must be because you've so inflated yourself
with the idea that people need your scathing comments and your complete lack
of respect to help them "learn."

Let me explain something that your parents should have explained to you long
ago. All your knowledge and experience is absolutely useless in improving
either yourself or the community if you only use it for self-justification
or to try to make someone else feel stupid. I came to Usenet expecting to
find intelligent discourse, and I did. I also expected to find many people
who didn't know a thing wandering around, and I did. What I wasn't expecting
(and I really should have) was an obviously educated man making a complete
fool of himself because he lacks the ability to reign in his temper, and is
inflated with a sense of self-importance far outweighing his worth to the
community.

1: he *can* answer hard CL and program design questions

2: he does

3: He takes the approach that "here is what you need to think
about to understand why you went wrong".  This is good, but it takes
work on the recipients part to use this.

4: he uses profanity well.

5: he gives out good book recommendations.

6: he has helped me in the past.

7: I have adoped his style of argument, and with people who
use at least half of their brain it generaly works.

8: he has high standards, I think that is good

In short he is someone who has proven himself to be a very handy
person to have around from a purly pratical point of view and he has
many qualities that I feel the world would be a better place if more
people had them.

good nignt

marc

Regards,
--
Nils Goesche
"Don't ask for whom the <CTRL-G> tolls."

  Good question.  Why do you remain and keep posting when you have previously
  been hurt and all you can do now is to hurt other people back, in some sort
  of private Middle East conflict?  What is _wrong_ with people like you?

  It used to be rewearding.  There are good people here.  Every once in a
  while, something _really_ rewarding happens.  This is generally called
  intermittent reinforcers if you ask a behavioral psychologist about it, and
  he will tell you they are much stronger than more constant rewarders.

  For instance, what I wrote about philosophy recently produced many very
  positive, and unexpected responses that I am still processing.  Helping
  newbies and whining kids is not rewarding.  Helping people with Common Lisp
  questions is not rewarding, anymore.  Talking about much more complex issues
  that underlie software design methodologies, learning strategies, etc, is
  rewarding.  This is not really the place to do that, but these are the people
  I have talked with and who have helped develop these ideas.  So sharing with
  them is rewarding.  Sharing anything with people like you is not.  Helping
  anyone who says he is "new to Lisp" and provides evidence of unwillingness to
  read a textbook on Common Lisp is an utter waste of time.  Half of them are
  retards like you and the other half never contribute to the forum when they
  understand something.

| Oh wait, I figured it out.  It must be because you've so inflated yourself
| with the idea that people need your scathing comments and your complete lack
| of respect to help them "learn."

  Some people actually respect me simply _because_ I help them, so they are
  never in need of scathing comments or anything else back.  Nor do they act
  with extremely paranoid self-defense when they are criticized for something
  they have, in fact, done.  If they do not understand it, they think about it
  and ask for assistance in understanding it.  These are mature adults.

  Does that sound odd to you?  I guess it is unfathomable to you at this point.

  Smart people have a _purpose_ when they ask a question.  They want to learn.
  You do not.  Your _purpose_ is to feel good about yourself.  As I said, I am
  sorry that I did not recognize that you suffered from an arrested development
  and still need to be hugged and kissed you when someone tells you that you
  have done something wrong.  I would have avoided answering you if I had known
  you would become such a hostile little cretin in return.  And since you have
  chosen to seek revenge rather than just state how you felt, you have turned
  into an aggressor that I will need to fight back.  Just how dumb can you be?

| Let me explain something that your parents should have explained to you long
| ago.

  Now I wonder what your parents were like.  They never let you grow up to
  become a human being.  Your behavior is more like that of a dog, which needs
  lots of hugging and stuff when it has done something wrong.  Lots of young
  people today behave like they have grown up like their parents' pets, and
  completely lack the intelligence to understand when to take care of their own
  emotional responses.  How you feel, Paul D. Lathrop, is actually _nobody_
  else's business.  Attempting to exact revenge for your emotional responses is
  just about the _least_ intelligent a human being can do, but it is somehow
  rewarded in some cultures and apparently by some parents such as yours.

| In short, if you don't like it here, don't let us peasants keep you from
| leaving.  Feel free to let the door hit you in the ass on the way out.

"Nils Goesche" <car...@cartan.de> ha scritto nel messaggio news:lkd6ubfsd7.fsf@pc022.bln.elmeg.de...

Regards,
--
Nils Goesche
"Don't ask for whom the <CTRL-G> tolls."

I am looking for professionalism. Erik has shown a decided lack of it.

That is all.

From what I read in your 'apology' it looks more like another attack
against Erik.  You are implying that Erik is not capable or willing to
meet the high standard that you hold your self to.  What has yet to be
demonstrated is are they a good thing for people to emulate.  Lets
look at what has happened so far:

1: You attack Erik.
2: you get caught
3: in your 'apology' you attack Erik again, by implication this time.
This is quite a bit worse then the original direct attack.

So please explain to me how you are someone we want around?  I can
not think of any reasons.

marc

  The simplest reason of all for why people leave USENET is the meta-debate.

  Nowhere outside of USENET do we find more bickering about the forum itself or
  the participants or language or whatever.  Is it "human" to do so?  I think
  not.  It is _stupid_ to do so, because every time someone has a personal
  coping problem that he needs to "share", like in some virtual group-hug, the
  signal-to-noise ratio drops.  Tehnical discussions however heated usually
  contain enough signal that the signal-to-noise ratio remains unchanged from
  normal discussion, and they are usually fairly quickly resolved as long as
  the technical is what matters.

  Everywhere I have been in real-life fora, I have seen a pattern: If people
  are focused on what they are doing with their time and maintain a clear sense
  of purpose, they develop a friendly atmosphere where helping each other reach
  their common and often also personal goals facilitates that atmosphere, from
  teamwork to large political bodies.  This is how USENET started.  But as soon
  as someone develops a different agenda, usually related to personal prestige
  or some other stupid emotional needs, it goes all animal.  For some bizarre
  reason, a behavioral pattern evolves that you do not find in animals who are
  fighting for something useful like food, but instead for their "rank" in some
  group.  As rational beings, they will make absolutely _no_ use of the result
  in normal discourse, but once they turn animal the next time, they seem to
  remember their "rank" and seek to do something about it.  My cat and I have
  had these fights over the years, especially when I have brought home another
  female that threatens to outrank her.  Cats tend to learn, but dogs will try
  again and again in subtle ways to check if their leader is still strong
  enough, and if a dog believes itself to be able to challenge the leader of
  the pack, it will just have to be killed if you do not want to fight it on
  dog terms.  Soem people on USENET behave in ways that suggest that the adage
  "on the Internet, nobody knows you're a dog" was not a joke.

  The solution to these problems is very simple: _Focus_ on what you came here
  to do.  If you came to ask for help because you are stuck, stay focused on
  that purpose.  If you feel bad because something you thought you had grasped
  slipped out of reach, it is nobody's business.  If you feel stupid because
  something you said was indeed stupid and you should have known better, you do
  _not_ help by complaining that somebody should have ensured that you did not
  feel bad.  If you need a security blanket, USENET is probably not for you.

  So why do we have these infertile meta-debates?  My guess is that some people
  have some sort of need to Do Something about perceived Unfairness.  That they
  are themselves usually grossly unfair does not bother them at all, of course,
  unless someone calls them on it, in which case they go insane and actually
  argue that they are in their right and the victim cannot even defend himself
  from their unfair, false, and misleading accusations.  People who react this
  way was the reason that more intelligent people fought wars to establish the
  rule of law and due process.  It is precisely to prevent the abuse of power
  by the emotionally disturbed and morally outraged that modern society evolved
  and developed mechanisms to protect suspected terrorists from incarceration
  and inhuman treatment, loss of freedom, torture, etc, just because some mad
  and angry child in power has lost his marbles, for instance.  The importance
  of a legal framework is frequently underestimated by people who think the
  police and the government are somehow their enemies, and so they discard the
  very mechanism they could have used to improve their condition, and regress
  to some animal form of revenge, instead of justice, where their own feelings
  guide their actions.  In the absence of rights and protecting authorities,
  every man is on his own defender of last resort.  However, human society has
  regulated justifiable defense since Roman Law ("moderamen inculpatæ tutelæ")
  and probably before that, and this regulation is quite necessary -- several
  large religions lack an upper bound on the amount of force allowable against
  some real or perceived aggressor.  Religion being the strongest expression of
  emotion, the passion and fervor of the moral outrage must be curbed lest it
  destroy the innocent and Lynch law rule.

  There is nothing worse than a group of people who manage to convince each
  other that what they are doing is proper and just.  Because of the strong
  elements of groupthink among even seemingly rational people, we poor, limited
  humans are in _very_ strong need of something that transcends the individual
  and curtails the expression of individual revenge responses when hurt.  We
  are _not_ a peaceful animal (even the cat that sleeps on my shoulders as I
  type is a murderer and unbelievably ferocious when some other female cat has
  invaded her territory, which is hard to believe if you visit a web site like
  http://www.blogjam.com/cute_little_kittens/).  We have historically survived
  by maintaining close-knit groups and adhering to the judgment of its leaders,
  but this is turning out to be the worst of all possible human traits as the
  group becomes too large for any one person to recognize and deal with, the
  number of judgments too high for any one person to internalize, the amount of
  new information too much for any one person to process -- groupthink is what
  holds us back as a species, severely limiting our ability to discover new and
  relevant truth.  Yet, even the very concept of majority in a democracy has
  been challenged when a sizeable group of people make obviously wrong choices:
  Jörg Haider, Jean-Marie Le Pen, George W. Bush, all have been challenged and
  resisted by people who knew better -- the frail consensus upon which society
  rests its trust in itself may have been hurt, but the situation would have
  been worse had nobody resisted and challenged these bad choices.

  My conjecture here is that these equally tireless and fruitless meta-debates
  are attempts to obtain a consensus on some principles of operation, lacking
  any authority has the physical power to hurt people if they disobey them.  I
  maintain that society has evolved into a relatively peaceful rule of law
  because we have found a sufficiently large club to hit people with and have
  found something that people actually agree on sufficiently that we can stand
  by and let the authorities hit people with said club.  However, this has not
  evolved on the Net.  In this particular sense, it retains its anarchic nature
  and meta-debates indicate that some people object not only to what they see
  as hitting, but to the very existence of the club.  I believe the latter is
  more important than the former.  Many people rebel against the concept of
  authority as such, not to any particular authority, and have such a hard time
  accepting that somebody else is an authority that it becomes necessary for
  them to _discuss_ this.  Like dogs who have to test the strength and power of
  their perceived pack leader, people who think in such terms behave in similar
  ways, even though there is no pack and no leader.  For such people, the waste
  that is the meta-debate is an important aspect to their existence -- because
  they think in terms of "rank", they also have to establish it if it differs
  from their expectations, and the more inflated their own expectations, the
  more they have to fight.

  This line completely invalidates any form of apology you might offer.