tobybailey

309 Reputation

6 Badges

17 years, 143 days

MaplePrimes Activity


These are replies submitted by tobybailey

We have had a site license for several years.  For a lot of our teaching we are switching to publishers online testing, mainly because TA is so poor.  We may well jump ship completely.

I completely agree about the LaTeX issues.  I would add that you can't put a less than symbol in the algorithm, or you couldn't last time I used it.  It seems to do some incredibly inept syntax checking of algorthms that flags all sorts of sensible use as errors but it passes all sorts of meaningless junk.  If it can't properly check the algorithms, why not just pass them through? I could go on.  And I completely agree that LaTeX authoring is or ought to be much quicker than using the dreadful MS equation editor.

And then the whole process of using the LaTeX converter is such a nightmare of clicking, uploading and downloading. (And the mot recent version has this error connected with inserting spaces all over the place). A grad student here wrote a python script to automate the whole process - why on earth hasn't Maple managed to produce some little application to do this automatically on the main platforms?

Agree to about maths rendering - things like primes are almost invisible.  And it doesn't scale which is a huge accessibility problem.

And then there is the "previewer" for formula input which when you preview 2*(a+b+c+d) produces 2(((a+b)+c)+d) which confuses students no end.

I could go on, but I think a clear example of the lack of seriousness of Maplesoft on this is that we are at Version 7 and we STILL have the following nonsense in the algorithm generation:

$c = frac(-1,-1);    produces (I kid you not)   $c = --1  AND this then causes an error if passed to Maple.

and

$p=0;

$q=$p;

produces $q= -0.

I mean really!  My guess is that the LaTeX converter and the underlying engine were inherited by Maple from whatever the preceeding system was that they took over, and nobody at Maple has actually understood how either of them works well enough to fix the glaring errors.

So I too am longing to hear that all this nonsense is going to be fixed.

 

Carl,

Many thanks.

This thought had ocured to me.    But now you repeat it, it seems reasonable.   So I have gone in, guessed which admin it was, reopened the assignment by changing its end time,   cured the problem and changed the end time back again.  I must say, it's a bit of a palavah!

Regards,

Toby

 

Carl,

Many thanks.

This thought had ocured to me.    But now you repeat it, it seems reasonable.   So I have gone in, guessed which admin it was, reopened the assignment by changing its end time,   cured the problem and changed the end time back again.  I must say, it's a bit of a palavah!

Regards,

Toby

 

Well, we (George  above and I are both in the same Department) have not gone quite that far.     Perhaps it would be wise to use only Maple for doing arithmetic, although one could worry about the time overheads.

Idon't think it is necessary to do things like $a = maple("0") though - as far as I can tell the maple returns are simply treated as strngs and I guess that this is completely equivalent to $a="0".  

Toby  

Well, we (George  above and I are both in the same Department) have not gone quite that far.     Perhaps it would be wise to use only Maple for doing arithmetic, although one could worry about the time overheads.

Idon't think it is necessary to do things like $a = maple("0") though - as far as I can tell the maple returns are simply treated as strngs and I guess that this is completely equivalent to $a="0".  

Toby  

It's just the usual issue - in 1D input they need escaping for the obvious reason of ambiguity:

  `\`x\``;

really does produce `x` as a name.

 

In 2_D input I have no idea what's going on and don't much care - I can't see any use for 2D input myself.

 

Toby

 

Interesting, but I was initially confused by your wording.   I thought for a minute you were claiming your modification itself changed O(n^2) to O(n) rather than what I assume you mean which is that using tables in some way is O(n) as opposed to incrementally building a set which is O(n^2).

By the way, if one can easily get some sort of upper bound for the number of entries, is it any quicker to use an Array rather than the table?

 

 

I thought that might be what you meant but I wasn't sure.

Toby

this seems to be just another version of this behaviour.    I still don't know whether I am misunderstanding something or whether this is a bug in evalf.

Certainly one does not get analogous behaviour with other indexable functions (if that's the word).  For example,

A := [3,map];
A[2](`+`,[1,2,3],5);

     [6,7,8]

and (correctly it seems to me)

A := [3,map];

A[2][2](`+`,[1,2,3],5);

  [1,2,3]+5


 

but I don't understand that either!    Specially  given

f:=[something,evalf]:

f[2][10](2.2222);

       2.2222

and for that matter also

f:=[something,evalf]:

g := f[2];

g(2.2222);

         2.2222

There is a danger in identify.   For instance,

x := evalf( sqrt(2) - 3/sqrt(7) );

identify(x);

returns

exp(  -(7/4) zeta(3) + (6/5) ln(2)  )

 

 

 

 

 

 

I doubt whether there is a way of posing this question to Maple in the form you state it, since Maple is just an algorithmic calculating machine.  Somebody somewhere may have written a "geometry package" that would allow you to pose the question to Maple in a way more or less equivalent to what you have stated, but that would just mean that somebody had done the work of formulating an algorithm for us.

What you would like seems essentially to be a machine capable of mathematical reasoning.  AI departments have people thinking about this sort of thing.  My impression is that they probably have not reached this far - or if they have it is only in the context of something for solving a particular range of geometric problems.

Of course, even if the AI people have created (or were to create) a mathematical reasoning program capable of solving your problem as you state it, ultimately it would be running on a computer and thus it would in a sense have the calculational method programmed in.  The question is, at what stage of complexity to you decide that the machine has decided for itself how to solve the problem.     One might take the view that choosing or seeing an appropriate method only makes sense for a creature (or machine(?)) with something like "free will".

 

 

 

Thanks.

 

Toby

Thanks.

 

Toby

Thanks for the historical summary. It explains a lot. I suppose in many ways my feeling that there should be a subs/eval acting precisely on the "op" structure is as much a desire for logical completeness as it is for any practical purpose. I guess for almost all practical uses they would agree and subs/eval as they are are more efficient.
1 2 3 4 Page 1 of 4