( Source: http://www.chris.com/ascii/index.php?art=objects/explosives )

On 08/02/2014 12:26 AM, Chris Cain wrote:
> On Friday, 1 August 2014 at 22:17:15 UTC, Timon Gehr wrote:
>> "contradictory assertions"
>>
>> "To what degree of certitude must the assertion be supported?"
>>
>> etc.
>>
>> I'll not go into more details because I actually intend not to
>> participate again in heated debate after already having shown strong
>> evidence.
>
> Frankly, I don't understand the point you're trying to make, so it's not
> really possible to rebut it.
>
> If your point is that assertions need to have evidence to be assertions,
> an easy counterproof of this is in your own quote: "To what degree of
> certitude must the assertion be supported?" ... It wouldn't have been
> called an assertion in that sentence unless it's something that is sure
> that it had been supported sufficiently (it would have been worded more
> along the lines of "To what degree of certitude must the assertion
> *candidate* be supported?")
> ...

Look, this is a really poor piece of logical reasoning, and for today I am expressedly fed up with arguing against obvious non sequiturs and immediately summoning new ones. In any case, my easy way out is to note that this is not my assertion.

To be really clear: My assertion was:

It is not true that anyone who accepts the "assertion definition" as the single conclusive way to give meaning to the term 'assert' will be able to use it to distinguish the meaning given to it by Walter and the other meaning under discussion, given to it by other notable people, in such a way as to find the former sensible and the latter nonsensical.

Sigh, I got involved again.

> I think it's clear that the definition of assertion being "statement of
> fact or belief" is compatible with the usage in that wikipedia article.
> ...

Yes. Please go back and read the relevant posts, find out what they assert and agree with them or argue against an assertion that was actually made. I have seen assertions such as: "the definition of the word really makes it obvious what the feature was originally intended to convey."

> In particular, "contradictory 'statements of fact'"

http://en.wikipedia.org/wiki/Fact

> [...]  actually make perfect sense.

Contradictory facts make perfect sense. Contradictory facts make no sense at all. Those two assertions expressed facts, unless this third assertion is wrong. The next assertion is wrong. The preceding assertions are statements of fact.

static if(!is(typeof(x))) int x;

...
     ...
         .
          .
           .


                             ____
                     __,-~~/~    `---.
                   _/_,---(      ,    )
               __ /        <    /   )  \___
- ------===;;;'====------------------===;;;===----- -  -
                  \/  ~"~"~"~"~"~\~"~)~"/
                  (_ (   \  (     >    \)
                   \_( _ <         >_>'
                      ~ `-i' ::>|--"
                          I;|.|.|
                         <|i::|i|`.
                        (` ^'"`-' ")

On 8/1/14, 3:05 PM, Daniel Gibson wrote:
> Yeah, it seems like assertion doesn't *have* to mean something like
> "promise", but can also be used in the sense of "claim" or "thesis",
> which is much weaker.

That said I'm perversely excited by the notion of a keyword called "thesis"... -- Andrei

On 08/02/2014 01:16 AM, Andrei Alexandrescu wrote:
> On 8/1/14, 3:05 PM, Daniel Gibson wrote:
>> Yeah, it seems like assertion doesn't *have* to mean something like
>> "promise", but can also be used in the sense of "claim" or "thesis",
>> which is much weaker.
>
> That said I'm perversely excited by the notion of a keyword called
> "thesis"... -- Andrei
>

:D

Maybe you should contact Hoare in order to extend http://en.wikipedia.org/wiki/C._A._R._Hoare#Apologies_and_retractions

On Friday, 1 August 2014 at 23:11:17 UTC, Timon Gehr wrote:
> ( Source: http://www.chris.com/ascii/index.php?art=objects/explosives )
>
> On 08/02/2014 12:26 AM, Chris Cain wrote:
>> On Friday, 1 August 2014 at 22:17:15 UTC, Timon Gehr wrote:
>>> "contradictory assertions"
>>>
>>> "To what degree of certitude must the assertion be supported?"
>>>
>>> etc.
>>>
>>> I'll not go into more details because I actually intend not to
>>> participate again in heated debate after already having shown strong
>>> evidence.
>>
>> Frankly, I don't understand the point you're trying to make, so it's not
>> really possible to rebut it.
>>
>> If your point is that assertions need to have evidence to be assertions,
>> an easy counterproof of this is in your own quote: "To what degree of
>> certitude must the assertion be supported?" ... It wouldn't have been
>> called an assertion in that sentence unless it's something that is sure
>> that it had been supported sufficiently (it would have been worded more
>> along the lines of "To what degree of certitude must the assertion
>> *candidate* be supported?")
>> ...
>
> Look, this is a really poor piece of logical reasoning, and for today I am expressedly fed up with arguing against obvious non sequiturs and immediately summoning new ones. In any case, my easy way out is to note that this is not my assertion.

I don't disagree that having to come up with an argument for you is poor. However, by not stating what it is you have provided "strong evidence" for, what does that leave me to do? I can simply say nothing but "what are you saying" but I had said that once already. It's a complete waste of time to argue with someone who refuses to make a clear position.

Frankly, if you're going to criticize others for making a "poor piece of logical reasoning", you should at least make something beginning to resemble a piece of logical reasoning yourself.

No, before you say anything, presenting evidence for a position you haven't stated isn't valid logical reasoning.

>
> To be really clear: My assertion was:
>
> It is not true that anyone who accepts the "assertion definition" as the single conclusive way to give meaning to the term 'assert' will be able to use it to distinguish the meaning given to it by Walter and the other meaning under discussion, given to it by other notable people, in such a way as to find the former sensible and the latter nonsensical.

And by what evidence do you back that up with? I think the only reasonable, logical thing to do when you state a fact to a computer program is that it accepts it as a fact. So the definition of assertion very much clarifies exactly the behavior described of assert. The only thing that differs from the definition is that the compiler will insert checks for your assertion under certain circumstances, but by the definition it would not be under obligation to do so (although, the fact that it's checking your assertions when you're debugging is infinitely useful, so it makes a lot of sense for it to go against the definition slightly for pragmatism).

>
> Sigh, I got involved again.
>
>> I think it's clear that the definition of assertion being "statement of
>> fact or belief" is compatible with the usage in that wikipedia article.
>> ...
>
> Yes. Please go back and read the relevant posts, find out what they assert and agree with them or argue against an assertion that was actually made. I have seen assertions such as: "the definition of the word really makes it obvious what the feature was originally intended to convey."
>
>> In particular, "contradictory 'statements of fact'"
>
> http://en.wikipedia.org/wiki/Fact

I hope you don't think that "state a fact or belief confidently and forcefully" and "fact" are necessarily the same thing. If it were, then the better definition of assert would have simply been "fact".

>
>> [...]  actually make perfect sense.
>
> Contradictory facts make perfect sense. Contradictory facts make no sense at all. Those two assertions expressed facts, unless this third assertion is wrong. The next assertion is wrong. The preceding assertions are statements of fact.

Contradictory STATEMENTS makes perfect sense. Since assertions are STATEMENTS of facts (or beliefs, to be clear), your entire argument is broken because you've used strawman.

It makes perfect sense to make contradictory assertions. Of course, your program failing to behave in a defined manner because of it (since you're, in effect, god of your program and the assertions you make will really be followed unlike people who might question your assertions and demand evidence) should be expected, just like if you wrote an incorrect if statement you can expect your program to have buggy behavior.

In particular, you cannot really reasonably expect that a program with incorrect statements (asserts, ifs, function call, whatever) to behave correctly. Writing an incorrect assert will (not surprisingly) result in a bug in a program just as if you wrote the wrong if statements in a binary search algorithm.

On 08/02/2014 03:11 AM, Chris Cain wrote:
> However, by not stating what it is you have provided "strong evidence" for,

Why would I need to? It is what you were arguing against: "You will notice it uses the word 'assertion' in a way that is incompatible with your claim that the "assert definition" rules out such an usage."

I think it should be clear from the context, no? Am I missing something?

> what does that leave me to do?

You can say that you still disagree but are in no position to argue further and we'll put aside the discussion. Mind you, I don't actually gain anything by convincing you, the investments have already been immense, and it would be a lot harder now since I appear to have insulted you, for which I want to express an apology.

> I think the only reasonable, logical thing to do when you state a fact to a computer program is that it accepts it as a fact.

Indeed. Note that actually we have defined an assertion to be a particular kind of statement of a fact _or belief_, ...

> So the definition of assertion very much clarifies exactly the behavior described of assert.

hence this conclusion is reached by a non-exhaustive case analysis on what kind of thing the assertion actually describes.

>
> Contradictory STATEMENTS makes perfect sense.

Yes, without further qualification, two statements may contradict each other. However, if we are e.g. talking about two statements that do not contradict each other, it would not make sense to claim them to be contradictory. There are many other qualifications that will imply that two statements do not contradict each other. The mere fact that we have two classes of STATEMENTS does not allow us to conclude that it makes sense to assume that their instances may be contradictory.

> Since assertions are

The same reservations as above apply.

> STATEMENTS of facts.

A statement of FACT is a statement that expresses a fact. I.e. there is an existing fact, and the statement expresses this fact. There is a difference between a fact and a statement of that fact, but if two statements of fact contradict each other, you are still left with two contradictory facts.

> your entire argument is broken because you've used strawman.

It is not a straw man even when assuming I actually claimed that you said contradictory facts make sense and then argued against that alternative claim because this claim you appear to think I argued against is an immediate consequence of the claim you made and which I wanted to refute.

On Saturday, 2 August 2014 at 01:55:49 UTC, Timon Gehr wrote:
> On 08/02/2014 03:11 AM, Chris Cain wrote:
>> However, by not stating what it is you have provided "strong evidence" for,
>
> Why would I need to? It is what you were arguing against: "You will notice it uses the word 'assertion' in a way that is incompatible with your claim that the "assert definition" rules out such an usage."
>
> I think it should be clear from the context, no? Am I missing something?

In one post you claim:
> In any case, my easy way out is to note that this is not my assertion.

In the next you claim I *was* arguing against what you were asserting.

I think the point about you lacking clarity makes itself absolutely crystal clear here.

Frankly, it's too difficult to understand you and I'm tired because I've had to argue with another guy about the definition of "test" because "That's not a test, that's code!" *facepalm* I really don't want to get into that sort of argument here again.

Frankly, I feel the thing you're going through is simply cognitive dissonance. I started this topic out thinking Walter was wrong but his points are far stronger and make a lot more sense when you consider more of the context behind `assert`, including the English definition (but also including the articles on assertions). So FWIW, I understand how you feel. But you're going to have to get over it and incorporate the new knowledge by arguing with yourself on the matter. Frankly, you're the person with the most to gain by casting aside misunderstandings of concepts. I don't gain anything by convincing you, so it makes more sense for you to argue with yourself, in this case. Really, I wouldn't mind doing you the favor of convincing you, but I have no tolerance for people whose cognitive dissonance is strong enough to argue against basic definitions of words (which you can verify by simply typing into Google "assert definition"). Honestly it makes it impossible for us to communicate when you won't even agree to accept a definition of a word which is widely accepted by every dictionary I can find.

If you have a problem with the understanding of "assertion", Google the definition until you're satisfied with a mental model behind it. It makes absolute abundant sense of why it would mean what Walter suggests. I think you'll find it really hard to continue to disagree with 10 dictionaries, so hopefully it'll cure you of your cognitive dissonance. Of course, by some luck you might find one dictionary I haven't found that might support your position, but I hope for your sanity and personal growth that you realize that 10 to 1 definitions means there's far more evidence against that.

On 08/02/2014 04:28 AM, Chris Cain wrote:
> ...

I retract my apology.

Here, I'll do you the favor of giving you a few more Google results with hopes that you'll start developing a mental model behind what the definition of assertion is:

Google itself: "state a fact or belief confidently and forcefully"

http://dictionary.reference.com/browse/assert: "to state with assurance, confidence, or force; state strongly or positively; affirm"

http://www.merriam-webster.com/dictionary/assert: "to state (something) in a strong and definite way"

http://www.thefreedictionary.com/assert: "To state or express positively; affirm"

http://www.oxforddictionaries.com/us/definition/american_english/assert: "State a fact or belief confidently and forcefully"

http://www.macmillandictionary.com/us/dictionary/american/assert: "to state firmly that something is true"


As a bonus, Googling "statement of fact" in case that isn't clear to you what that means (maybe that's what's throwing you for a loop? You seem to be equating statement of fact and facts so you've clearly got some understanding wrong for the definition of those two):
http://en.wikipedia.org/wiki/False_statements_of_fact

False statement of fact... You can't possibly suggest that there exists false facts, but false statements of facts do exist, proving that statement of fact != fact.

On Saturday, 2 August 2014 at 02:32:45 UTC, Timon Gehr wrote:
> On 08/02/2014 04:28 AM, Chris Cain wrote:
>> ...
>
> I retract my apology.

Of course. The worst curse I could wish upon you is that you stick to your guns. So stick to your guns, my friend. :-)

On Friday, 1 August 2014 at 21:06:42 UTC, H. S. Teoh via
Digitalmars-d wrote:
> Well, if we could turn back the clock and redesign D based on our
> experience, we'd have "throwing" and "mutable" as modifiers instead of
> "nothrow" and "immutable". :)
>
> Not to mention "impure" instead of "pure".
>
>
> T

D3 anyone? :)