Actually, I am unhappy to be forced to use an overrestricted assumption like positive so as to get a simple result. It should work with real, and without a surgery.

I hope that this suggestion does not turn into another dissambiguation box  as with the assignment to f(x) (function definition vs remember table assignment).

I think that Jacques has posted on this issue one or two years ago (I do not have the link right now). If I remember correctly what he said, a parser alone cannot decide correctly in all situations, and some stronger "infrastructure" handling the proper context would be needed.

I hope that this suggestion does not turn into another dissambiguation box  as with the assignment to f(x) (function definition vs remember table assignment).

I think that Jacques has posted on this issue one or two years ago (I do not have the link right now). If I remember correctly what he said, a parser alone cannot decide correctly in all situations, and some stronger "infrastructure" handling the proper context would be needed.

No, I did not mean whether you have a review, or will write it shortly, but just  when fit, hopefully before 30 years...

I could catch only part of the material in those slides because of the CS
jargon, but I was a bit surprised by the historical link Maple -> Sage as I
had thought that the main influence on Sage came from some other CAS.

The way you put the issue of weak vs strong types sounds a bit funny to me. By
those terms, in Physics research it is not know what is being done: at
exploring new ground, most likely, a strong underpinning mathematical theory
is missing.

And, as Physics deals mainly with analysis and geometry, it sounds that a
strong types will become, most frequently, a hindrance. Something like the
less typed the better.

This issue might help also to explain why I find Axiom so hard to understand
and make something useful with it, as it was developed by (and fits well into)
minds more familiar with fields like logic or algebra. I.e. "strong typed minds".

In this regard I have seen that a (qualitative/semiquantitative?) measure of the
"degree of typing" was sometimes posed, as in these articles (ref1, ref2).

And here and there I have seen that CAS were sorted after such a magnitude. E.g. this post suggests the following order:

Maple < Mupad < Axiom

If this order makes any sense, I wonder where other systems lay. Indeed I
was a bit surprised at reading in the Wikipedia article Comparison of programming languages that Mathematica has a strong
type strength (and no mention of Maple, by the way).

I cannot get you wrong as. I have no idea of what the abstract theories of
analysis, geometry and imperative programming are. Most likely, whatever I
have ever done are computations with specific objects.

I do not understand either when you talk about 'denotational semantics'. If you
mean something like stated in this article, I am perplexed. I read:

Broadly speaking, denotational semantics is concerned with finding mathematical objects that represent what programs do.

Al least for a mathematical system like Maple, this issue seems to me
upside-down. What I want is a computational representation of (some)
Mathematics (and as faithful as possible). What is the point of inventing a
mathematical representation of a routine for calculus that outputs
mathematical nonsense? (like several ones in Maple do).

I guess that what you say about a "decidable type system" is related to these
two articles (ref3, ref4). But I do not see why that 'apparent domain' is decidable while the 'effective domain' is not. Is there a proof for that?

I do not believe that generic polls like this one, could serve much if Maplesoft seeks for feedback in regards to development priorities in this new stage. So, I think, more focused polls are needed about interest on specific features or packages. And I mean visible results polls, unlike the recent user interface survey.

Have you tried, for instance,  the commands that I have posted above?

For help I have told you, and you have written it: type ?UserLib!!! .

Have you tried, for instance,  the commands that I have posted above?

For help I have told you, and you have written it: type ?UserLib!!! .

Indeed, spotty observation of this library's commands shows that part of its functionality is somehow available in newer releases of Maple. But I do not know how much.

Indeed, spotty observation of this library's commands shows that part of its functionality is somehow available in newer releases of Maple. But I do not know how much.

I have this library installed in Maple 10.06. I have not tried with Maple 13 yet. It is an assorted mix as it comes with top level commands like:

Release(); 
                                  10

and packages:

with(AlgLists);
                         [&*, &+, &-, &/, &^]

[a,b] &/ [1,2];

                               [a, b/2]

I have this library installed in Maple 10.06. I have not tried with Maple 13 yet. It is an assorted mix as it comes with top level commands like:

Release(); 
                                  10

and packages:

with(AlgLists);
                         [&*, &+, &-, &/, &^]

[a,b] &/ [1,2];

                               [a, b/2]

I think that this MathML-like stuff is undocumented. See here for a simple example of MathML.

I think that this MathML-like stuff is undocumented. See here for a simple example of MathML.

If part of the Maple system routines are in binary format then the dependency tracing should use suitable tools for them.

I am not sure what do you mean by "normal Maple usage". I think that there are many different patterns of Maple usage nowadays. Just reading e.g. ?define_external or ?OpenMaple,VB,API I find things about compiled code and other issues with which I am not familiar, but may be normal for some other user.

I have little idea how these things work, but the existence of an application like the Dependency Walker, suggest to me that scanning these dependences might not be impossible.