The situation is even more curious. On one hand, I have missed two names with property 'LinearMap' : 'limit' and 'Limit' . So they are also 'continuous'...

On the other hand, for the property 'monotonic':

about(monotonic);
monotonic:
  a function which over the reals and where defined is non-decreasing (or non-increasing)
  a known property having {mapping} as immediate parents
  and {StrictlyMonotonic} as immediate children.

It holds in the property lattice:

`property/ChildTable`[monotonic];
`property/ChildTable`[op(%)];
                             {StrictlyMonotonic}
                                 {LinearMap}

Hence, for these 20 names:

[hankel ,mellin ,fouriersin ,Diff , invztrans , invfourier ,
invlaplace , hilbert , Int, fourier , Sum, ztrans, 
fouriercos, laplace, diff, invhilbert , sum, int,limit,Limit]:
map(is,%,monotonic);

[true, true, true, true, true, true, true, true, true, true, true, true, true,
  true, true, true, true, true, true, true]

What could be argued about that?

As the commands 'Int and 'int' are typically intended to represent integration of functions on the reals, and that is the "definition" of 'continuous', let us focus on such examples.

And these commands represent indefinite integration as well as definite integration on both bounded and unbounded intervals.

Begining with the question when indefinite integration is continuous, I was thinking  as first example on diverging (unbounded) functions like 1/x^n with n integer >=2. These functions are not continuous for every real value of x (ie divergent as x->0 and undefined as real functions at x=0). The same for their integral. Idem for any linear combination.

It seems in this example that integration cannot be continuous in the sense of the definition, even being linear.

means that the integral is bounded?

under the change of variables x->X, t->T becomes a function of X, not T.

d:=c*diff(x(t),t) + g(x(t)) / t - 1 ;
convert(d,y_x);
                       d             c t(x)
                       -- t(x) = - -----------
                       dx          g(x) - t(x)

to simplify the analysis, eg like:

-(1/2)*Rs*(-2*Rs*Cs+t+2*exp(-t/(Rs*Cs))*Rs*Cs
+exp(-t/(Rs*Cs))*t)/(Lp*(-1+exp(-Pi*sqrt(Cp*Lp)/(Rs*Cs))
*exp(-t/(Rs*Cs))))=n;
subs(Rs=ts/Cs,t=tau*ts,(Cp*Lp)^(1/2)=alpha*ts,%);
collect(%,ts);
subs(ts^2=Cs*Lp/beta^2,%);

                -2 + tau + 2 exp(-tau) + exp(-tau) tau
           -1/2 -------------------------------------- = n
                                                    2
                (-1 + exp(-Pi alpha) exp(-tau)) beta

So, the adimensionalized time tau depends on two parameters: alpha and n*beta^2=gamma, say. You can make a 3D plot then.

to simplify the analysis, eg like:

-(1/2)*Rs*(-2*Rs*Cs+t+2*exp(-t/(Rs*Cs))*Rs*Cs
+exp(-t/(Rs*Cs))*t)/(Lp*(-1+exp(-Pi*sqrt(Cp*Lp)/(Rs*Cs))
*exp(-t/(Rs*Cs))))=n;
subs(Rs=ts/Cs,t=tau*ts,(Cp*Lp)^(1/2)=alpha*ts,%);
collect(%,ts);
subs(ts^2=Cs*Lp/beta^2,%);

                -2 + tau + 2 exp(-tau) + exp(-tau) tau
           -1/2 -------------------------------------- = n
                                                    2
                (-1 + exp(-Pi alpha) exp(-tau)) beta

So, the adimensionalized time tau depends on two parameters: alpha and n*beta^2=gamma, say. You can make a 3D plot then.

is 16-18 March, 2005. The interview was about history, not forecast. And the reference applies to 1995-2005. The next sentence does not imply a conection with the interview, but it is written in a way that an unattentive reader could be induced to think so.

is 16-18 March, 2005. The interview was about history, not forecast. And the reference applies to 1995-2005. The next sentence does not imply a conection with the interview, but it is written in a way that an unattentive reader could be induced to think so.

that I see (p.55) is:

At some point we were recovering market share from Mathematica or regaining the market share that we had before. That didn’t continue. That didn’t happen in ’95 and onwards. I think that that’s what pushed Maple farther and farther behind.

And this one (p.46):

If anything, in the battle with Mathematica, our biggest mistake was that one, was to concentrate on efficiency and correctness in algorithms and coverage, as opposed to how easy it is to input information into the system and how pretty and convenient is the output.

that I see (p.55) is:

At some point we were recovering market share from Mathematica or regaining the market share that we had before. That didn’t continue. That didn’t happen in ’95 and onwards. I think that that’s what pushed Maple farther and farther behind.

And this one (p.46):

If anything, in the battle with Mathematica, our biggest mistake was that one, was to concentrate on efficiency and correctness in algorithms and coverage, as opposed to how easy it is to input information into the system and how pretty and convenient is the output.

In this interview (p. 33) he says:

At some point, we were not distributing the library, but I had made sure that you could print any function from Maple. Well, you would get it without the comments, but at least you’d get the function.

It is not explicit in this paragraph, but it sounds as a change made when  Watcom began to distribute Maple, or close to that.

It is quite curious, because in the previous answer he seems to praise the benefits of distributing the library code:

It also made the code transparent, so we had lots of people that were users of Maple that would say, “This function is very nice, but you have a bug here. Fix it,” and they would even send us the fix. That was because it was in the libraries, because these were
people that needed the math done, and if it was not doing it, they would go in and fix/improve it. They had all the tools to solve it—they knew the language and they had the code, so they could
go and fix it.

In this interview (p. 33) he says:

At some point, we were not distributing the library, but I had made sure that you could print any function from Maple. Well, you would get it without the comments, but at least you’d get the function.

It is not explicit in this paragraph, but it sounds as a change made when  Watcom began to distribute Maple, or close to that.

It is quite curious, because in the previous answer he seems to praise the benefits of distributing the library code:

It also made the code transparent, so we had lots of people that were users of Maple that would say, “This function is very nice, but you have a bug here. Fix it,” and they would even send us the fix. That was because it was in the libraries, because these were
people that needed the math done, and if it was not doing it, they would go in and fix/improve it. They had all the tools to solve it—they knew the language and they had the code, so they could
go and fix it.

It is not my field, but I know that people do Monte Carlo integration for multidimensional integrals.

It is not my field, but I know that people do Monte Carlo integration for multidimensional integrals.

Differentialgleichungen (in German) is a classic reference. It has a section like a phone directory where lo look for ODEs and their known solutions.