Your code deals with the (much easier) problem of finding contiguous increasing sub-sequences.  I meant an arbitrary sub-sequence.

with r defined as just sin(v^2)+cos(v^2)*v^2 (the numerator) under the assumption v>0 , we get

is(0 < r), is(0 <= r);
                             false, true

where the second one is clearly a bug.  In fact, as far as I can tell, the bug is in `is/RealRange` the line

`property/PropInclusion`(objprop,newprop)

has things backwards, ie it should really be

`property/PropInclusion`(newprop,objprop)

If I am right, I wonder how many bugs in is (and thus signum, etc) are due to this?

with r defined as just sin(v^2)+cos(v^2)*v^2 (the numerator) under the assumption v>0 , we get

is(0 < r), is(0 <= r);
                             false, true

where the second one is clearly a bug.  In fact, as far as I can tell, the bug is in `is/RealRange` the line

`property/PropInclusion`(objprop,newprop)

has things backwards, ie it should really be

`property/PropInclusion`(newprop,objprop)

If I am right, I wonder how many bugs in is (and thus signum, etc) are due to this?

whenever I went to conferences, and still to this day, as soon as people learn that I have some connection to Maple, they unload their personal cache of bugs (verbally) onto me.  And they are often quite frustrated as they do this.

I wonder why Derive never caught on? 

whenever I went to conferences, and still to this day, as soon as people learn that I have some connection to Maple, they unload their personal cache of bugs (verbally) onto me.  And they are often quite frustrated as they do this.

I wonder why Derive never caught on? 

Recent enough Maple users have been pushed towards Embedded Components, and older users write their Maplets 'by hand'.  If the Maplet Builder "doesn't work", a lot of people might have just decided to avoid it rather than bother reporting problems.

As a matter of fact, my experience is that less than 10% of Maple users bother complaining about the problems they encounter.  When I was a manager at Maplesoft, I found this extremely frustrating as it meant that I had just a fraction of the data I needed to figure out where to concentrate my efforts!

Recent enough Maple users have been pushed towards Embedded Components, and older users write their Maplets 'by hand'.  If the Maplet Builder "doesn't work", a lot of people might have just decided to avoid it rather than bother reporting problems.

As a matter of fact, my experience is that less than 10% of Maple users bother complaining about the problems they encounter.  When I was a manager at Maplesoft, I found this extremely frustrating as it meant that I had just a fraction of the data I needed to figure out where to concentrate my efforts!

Unfortunately for these kinds of issues, Maplesoft's Technical Support really are the best source of information.  You'll have to wait until Monday morning though...

What you describe is most definitely unexpected behaviour, so I would call it a bug.  You should submit it (see the Submit Maple Software Change Request menu on the left-hand-side).

You make a splendid point in reference to end-users vs developers re: pointy-clicky interfaces.  I couldn't agree more.

These kinds of GUI builders are so often really awkward - I think I only saw the Maplet builder once, and saw it as yet another example of an awkward builder for a task that was better and more easily done by a few lines of text.  I still laugh when I hear ardent Mac users sing the praises of Quicksilver!  If only Maplesoft went through the same epiphany as those users...

[I helped with this particular code, so I know it fairly well, at least the version in Maple 8; I don't think it has changed that much since]

Yes, something equivalent is implemented in MathML[ExportPresentation], but it was not written to be re-used in other contexts. In other words, I would rather doubt if the pieces are available in any other way than as a model to copy.

That's too bad, because the CodeGeneration package is written differently, and is explicitly built so that its front-end processor can be re-used and matched with specific back-ends.

Congrats on the 500.  Very thoughtful post, I do hope that the 'right' people will read it.

Myself, I never understood why Embedded Components were introduced into Maple.  I mean, why couldn't Maplets have been improved to have 2 modes and act as both dialogs and be embedded into the worksheet?  Maplets have a very nice programmatic interface that is very convenient to use when you're doing anything serious.  As you found out, doing something 'serious' with an interface that requires clicking is not worth it (200 clicks?!?).  The better design would have been to create a GUI interface to creating Maplets for those cases you wanted a simple component embedded into the worksheet.

Note that I was quite careful to say "almost always" when talking about finding gaps.  Yes, one can construct matrices with no gaps.  However, it seems that for naturally occuring phenomenon with real information in them, there is usually a gap [counter-examples here would be very interesting].

So yes, it would be very interesting to see rank-3 vs rank-4, as it seems that that is where a lot of information "finally" comes in, and that going up to 7 gives you the majority of the information, and the rest just fills-in the details.

It would be quite difficult to single-out a user and have some notifications not sent!  I would first suspect that there might be a problem with the email address that you have registered with the site.