Re: Maple's "fd" crash code, and FORTRAN in Maple


On Wed, 22 Jun 2005 15:25:12 +0200, Christopher Creutzig
<christopher@xxxxxxxxxxx> wrote:

>Craig Carey wrote:
>
>> A MuPad person (Mr Christopher Creutzig), said earlier at sci.math.symbolic,
>> that some of Maple was in FORTRAN.
>
> Did I? I can't remember knowing or saying something to this end, and
>groups.google.com only turns up some speculation of mine in 2003 that
>Maple *might* include some Fortran, which was based on the cooperation
>with NAG.
>

I guess I got the quote right. Anyway the percentage ws not indicated.

Can you get the percentage of FORTRAN in Maple ?. Why doesn't Maplesoft
monitor the sci.math.symbolic group and visit and assert that Maple does
not contain FORTRAN?.


>> | unknown software exception (0xc00000fd) at 0x00245713.
....
> Blame Microsoft. At the point where the message is printed, Maple is
>no longer running, the message is generated by the operating system.
>According to msdn, the fd is ?STATUS_STACK_OVERFLOW?, usually caused by
>a recursion nested too deeply.


There is not an explanation here of the big variability in time, for the
crash to appear. Excessive recursion can caused by a bug and not only by the
input data.

Windows 2003 does not produce 2 hex numbers and memory runs low. So Maple
too, can catch the exception and produce something better than 2
hexadecimal number.

A quick estimate is that Maple should not cost more than about US$15.

That "fd" hints at giving up crisis inside of Maplesoft.
Maybe that alone causes bugs to be left uncorrected.

----

Here is my list some categories thatr the source code of symbolic algebra
solver software, can be allocated to. Debugging bugs in category 1 can be
swift.


* 1: Swapping & changing pieces of expression structures.

Also: the 3GL code supporting the user-side language.

* 2 intricate maths algorithms.

* 3 parser code [guessing and workarounds.]

* 4 GUI code [coding can require persistence and debugging]

* 5 bug-free memory management code. [Gets debugged and left alone]


Debugging category 2 problems might be very slow, occasionally.

Let me summarize a little:

(-) Mr B was almost saying that Maplesoft has a long queue (large set) of
bugs,

(+) I say that most of the bugs reall can be removed rapidly. I.e. long
queues of Maple bugs could be shorted at a fast speed (or categories 1
and some of category 2).


---

Here is a character profile of a speculated-to-exist Maplesoft programmer..

+ Mr Apple Bonfire (of MapleCore) loses hope on 3GL bugs long before
+ other programmers would.
+
+ Bug removal is a project, and it can be left unfinished over years (just
+ like placing 3,000 small cardboard boxes against an office wall).
+
+ Mr Bonfire's enthusiasm can falter in maybe 40 minutes. That is quick
+ enough to seem suspect.
+
+ But maybe the company views debugging as a way to lose customers.

---

Here is a backtrace report from my Tope program.
It improves on the 2 hexadecimal "fd" numbers of the corporation.

The text of the call history is in a dark green colour:

--------------------------------------------------------------------------
Tope 21-June-2005

1: 3*(p or q);

SYSTEM.ASSERTIONS.ASSERT_FAILURE
Message: tp-m1.adb:1570
007216B1 in ada.exceptions.process_raise_exception at a-except.adb:1320
00723C07 in system.assertions.raise_assert_failure at s-assert.adb:46
0065D26F in tp.m1.debug.check at tp-m1.adb:1570
00655EB0 in tp.m1.g2.assign_fast at tp-m1.adb:156
00657504 in tp.m1.g2.replace_top_with_part_thunk at tp-m1.adb:356
0065732A in tp.m1.g2.b4_replace_top_with_part_delicately at tp-m1.adb:342
006570B6 in tp.m1.g2.b4_fixup_assume_booleans at tp-m1.adb:314
00663EF5 in tp.m2.b4_merge_associative at tp-m2.adb:523
0066D9CA in tp.rats.nums.multiply_ass_fast at tp-rats-nums.adb:242
006BCB14 in tope2_model.act at tp-model2.adb:1846

2:
--------------------------------------------------------------------------

A rising topic might be:
their efficiency at successfully closing incoming bug reports about Maple.





Craig Carey

.

Relevant Pages

  • Re: Maple bugs: Thomas Richard: Hurrah, Maple quality improves! - Example 4
    ... SB> proof that maple is getting worse. ... You can find such examples easily in Maple Bugs Encyclopaedia ... The Software Quality Research Laboratory's team conducts ... architect the GEMM and the VM machines, and at Cyber Tester ...
    (sci.math.symbolic)
  • Re: Maple bugs: Thomas Richard: Hurrah, Maple quality improves! - Example 4
    ... SB> proof that maple is getting worse. ... You can find such examples easily in Maple Bugs Encyclopaedia ... The Software Quality Research Laboratory's team conducts ... architect the GEMM and the VM machines, and at Cyber Tester ...
    (sci.math)
  • Re: Maples "fd" crash code, and FORTRAN in Maple
    ... Maybe a bit of Fortran lately. ... suggest that Maple contained FORTRAN. ... the Visual C compiler is eliminable. ... 2-3 bugs can be hit per hour. ...
    (sci.math.symbolic)
  • Re: A free copy of Maple or a frre maple webservre
    ... It is not so much that he 'discovered' the bugs but that he _generated_ the bugs. ... But you need to take what I say in the context I said it - mainly I was agreeing with Karen that you can't really verify results too well in closed source software like Maple. ... Of course, since Sage is open-source software, with many components been around a very long time, I expect some Maple developers have looked at some of the Sage code - some might even have written it. ...
    (sci.math.symbolic)
  • Re: A free copy of Maple or a frre maple webservre
    ... Vladimir Bondarenko is a skilled mathematician who has a keen understanding ... of the internals of Maple. ... It is not so much that he 'discovered' the bugs ... contribute to the evolution of Maple. ...
    (sci.math.symbolic)