William Stein declares, Mission Accomplished!

[rjf]

Rather than invading Afghanistan, Prof. Stein moves on to Purple
Sage,
for doing his own arithmetic geometry research.


http://389a.blogspot.com/2010/09/purple-sage.html

RJF

[Bill Hart]

Yawn, that's this morning's news. We got an oil platform on fire in
the gulf spilling oil this avo. Much more interesting.

Bill.

> --
> You received this message because you are subscribed to the Google Groups "sage-flame" group.
> To post to this group, send email to sage-...@googlegroups.com.
> To unsubscribe from this group, send email to sage-flame+...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/sage-flame?hl=en.
>
>

[kro...@uni-math.gwdg.de]

Citing from http://389a.blogspot.com/2010/09/purple-sage.html

The requirements of 100% test coverage, documentation, and peer review will be removed

 

And the development will fall  back to the previous behaviour pattern which will likely result in coding mess..

ha-ha !  ( sorry, couldn't resist, its sage-flame )



Jack

[Bill Hart]

That's not the main Sage project, but a research only project.

--
You received this message because you are subscribed to the Google Groups "sage-flame" group.

To unsubscribe from this group and stop receiving emails from it, send an email to sage-flame+...@googlegroups.com.

To post to this group, send email to sage-...@googlegroups.com.

Visit this group at http://groups.google.com/group/sage-flame.
For more options, visit https://groups.google.com/groups/opt_out.

[kro...@uni-math.gwdg.de]

Yes, Bill, I know that Purple Sage is not the main Sage project.

I suspect, without verification (e.g. by theory or independent implementations) the research results may not be worth a penny.

The same holds unfortunately for our package (in development) for resolution of singularities I'm working on together with Anne Frühbis-Krüger.
By applying various test strategies and reviewing code we found several bugs in the package itself and  in used routines from Singular.
And it is very likely that there are even more bugs not discovered yet.

[rjf]

On Thursday, January 16, 2014 11:51:03 AM UTC-8, kro...@uni-math.gwdg.de wrote:

Yes, Bill, I know that Purple Sage is not the main Sage project.

I suspect, without verification (e.g. by theory or independent implementations) the research results may not be worth a penny.

If the implication here is that Sage is, by contrast, free of bugs and worth money, eh.  It is certainly not free of bugs.

There are rather few places to sell theorems for money.

The same holds unfortunately for our package (in development) for resolution of singularities I'm working on together with Anne Frühbis-Krüger.
By applying various test strategies and reviewing code we found several bugs in the package itself and  in used routines from Singular.
And it is very likely that there are even more bugs not discovered yet.

 

There are ways of estimating the number of bugs not yet discovered, in case you are interested.

 

[Bill Hart]

On 16 January 2014 20:51, <kro...@uni-math.gwdg.de> wrote:

Yes, Bill, I know that Purple Sage is not the main Sage project.

I was just pointing out that this is sage-flame, not purple-sage-flame.

[Ondřej Čertík]

On Thu, Jan 16, 2014 at 1:23 PM, Bill Hart <goodwi...@googlemail.com> wrote:
>
>
>
> On 16 January 2014 20:51, <kro...@uni-math.gwdg.de> wrote:
>>
>>
>> Yes, Bill, I know that Purple Sage is not the main Sage project.
>
>
> I was just pointing out that this is sage-flame, not purple-sage-flame.

The mission is not accomplished unless you can do 10 pullups.

Ondrej

[Tom Boothby]

With a sprinkle of potassium chloride (a quite real salt that we call
'fake salt' in my family on account that my grandma uses it and we
ridicule her for... well, everything, really), we could make turn this
into purple-sage-flame.

[William Stein]

On Thu, Jan 16, 2014 at 6:57 PM, Tom Boothby <tomas....@gmail.com> wrote:
> With a sprinkle of potassium chloride (a quite real salt that we call
> 'fake salt' in my family on account that my grandma uses it and we
> ridicule her for... well, everything, really), we could make turn this
> into purple-sage-flame.

I did 10 handplants tonight. Here's two: http://youtu.be/UnGUAGeHWKc

William Stein
Professor of Mathematics
University of Washington
http://wstein.org

[Bill Hart]

Everyone knows that faceplants are no substitute for pull ups. But I shouldn't get too excited. I can only do about 6 pull ups.

[rjf]

Handplants and faceplants are different, though perhaps in a topological sense they are equivalent.

I don't see why you think this is on topic for sage-flame. Why not post it on some skateboarding
or William-Stein-athletic-achievements blog?

While people are reading this, I might make an on-topic comment.   I think that sage-devel has been
really low quality recently.  Typical posts are  dull, ill-informed or both. Maybe Sage is moving
 towards the heat death of maximum entropy, where all its bits will be random.

RJF

[Tom Boothby]

On Fri, Jan 17, 2014 at 8:49 AM, rjf <fat...@gmail.com> wrote:

> While people are reading this, I might make an on-topic comment. I think
> that sage-devel has been
> really low quality recently. Typical posts are dull, ill-informed or both.
> Maybe Sage is moving
> towards the heat death of maximum entropy, where all its bits will be
> random.

Perhaps you can make a contribution to Sage for once, and submit a
patch that does exactly that. I'd be more than happy to review it.

[rjf]

Every time there is a new version of Maxima there is, in effect a patch
to Sage.  If it is reviewed by Sage people, it is likely to be perfunctory
(I.e. it still loads and runs the previous tests).  So the review process is,
to that extent, a sham.
If I contribute to a new version of Maxima in some way, I contribute to Sage
as well, at least at the time that Sage catches up to that version of Maxima.

So your notion that I do not contribute to Sage is incorrect.  Your notion
that you (or someone) would have to "review" it is also off base.  You are
of course welcome to report bugs that you find, in some review process,
in Maxima to the appropriate forum.   Occasionally such reports appear.
Sometimes they have the form, roughly,  "I've neglected to read the
documentation for Maxima, and so have mis-used it in the following fashion
and wonder why the response is not what I anticipated I would get, and so
I report it as a bug."

Occasionally the bug reports are in fact reporting bugs, and for those
people are presumably grateful.

 RJF

[Tom Boothby]

Maxima is the worst possible thing for Sage. If there wasn't Maxima,
somebody would have reinvented the wheel. That wheel would fit into
the rest of Sage better than Maxima does. My suggested "maximum
entropy patch" may actually deal less damage than your past
contributions. One of these days, sympy will fit the bill and
hopefully we'll be able to forget Maxima's ghastly existence
altogether.

[rjf]

On Friday, January 17, 2014 9:57:03 AM UTC-8, Tom wrote:

Maxima is the worst possible thing for Sage.

Apparently the people in charge don't seem to take their marching orders from you, Tom.

 

 If there wasn't Maxima,
somebody would have reinvented the wheel.

The existence of Macsyma/Maxima has not prevented people from reinventing it.
The sad thing is that the reinventors tend to copy the bad along with the good,
and so there are instances of "system independent bugs"  which appear in
Macsyma, Maple, Mathematica, ... and probably in SymPy.
 

 That wheel would fit into
the rest of Sage better than Maxima does.

I think that the problem of "fitting" would have been alleviated if Sage had been
written in a more appropriate language.  But then I'm not sure how the analogy
really works.  Fitting wheels?
 

 My suggested "maximum
entropy patch" may actually deal less damage than your past
contributions.  One of these days, sympy will fit the bill

It is hard to make predictions, especially about the future.

 

and
hopefully we'll be able to forget Maxima's ghastly existence
altogether.

It is impossible for one to forget things that he is ignorant of.
So I expect that what you mean is that you would be happy to find
that your (and possibly others') laziness is ultimately redeemed
by the sympy development making knowledge of Maxima irrelevant.
I expect that you will also be ignorant of sympy development, and
so you would have a balanced perspective.  Symmetrically ignorant.

My expectation is that for a long time to come sympy will be
severely less capable than (say) Maple, Mathematica, Axiom, Maxima, ....
While it is possible that it will be a step forward in computer algebra /
symbolic mathematics systems, there is certainly no assurance of
that happening.  Even with smart people working hard on it.

If there is an academic or technical paper that explains how sympy
developers recognize the shortcomings of the design of previous CAS
and how they are overcoming them, I am not aware of such item.
Perhaps someone can point the way?

If Tom (and others) believe that the key to "doing it right" is to use
python instead of lisp for symbolic mathematics system building,
I strongly doubt it.  In fact, I read somewhere that the core of sympy was
rewritten in C.  That should keep it, what is the claim?   readable.

Incidentally, while a nice user interface is -- um, nice -- that does not
seem to me to be the key to , essentially,  storing and using all
symbolic mathematical knowledge. 

So in brief,  what's the story with sympy, anyway?

RJF

[Ondřej Čertík]

On Fri, Jan 17, 2014 at 12:41 AM, Bill Hart <goodwi...@googlemail.com> wrote:
> Everyone knows that faceplants are no substitute for pull ups. But I
> shouldn't get too excited. I can only do about 6 pull ups.

Hm, you can lift 2x more on bench press than I do, but I can do twice
as many pullups as you do?
How is that possible? I weight 210 lbs.

Ondrej

[Tom Boothby]

On Fri, Jan 17, 2014 at 11:10 AM, rjf <fat...@gmail.com> wrote:
>
>
> On Friday, January 17, 2014 9:57:03 AM UTC-8, Tom wrote:
>
> So in brief, what's the story with sympy, anyway?
>

Fuck if I know, captain. My knowledge is symmetric in all things
aside dinosaur-baiting.

[Bill Hart]

On 17 January 2014 20:49, Ondřej Čertík <ondrej...@gmail.com> wrote:

On Fri, Jan 17, 2014 at 12:41 AM, Bill Hart <goodwi...@googlemail.com> wrote:
> Everyone knows that faceplants are no substitute for pull ups. But I
> shouldn't get too excited. I can only do about 6 pull ups.

Hm, you can lift 2x more on bench press than I do, but I can do twice
as many pullups as you do?
How is that possible? I weight 210 lbs.

It's just whatever you practice. I used to not be able to do dips. It took me months before I could do one. A few weeks later I could do 12.

[Ondřej Čertík]

On Fri, Jan 17, 2014 at 2:28 PM, Bill Hart <goodwi...@googlemail.com> wrote:
>
>
> On 17 January 2014 20:49, Ondřej Čertík <ondrej...@gmail.com> wrote:
>>
>> On Fri, Jan 17, 2014 at 12:41 AM, Bill Hart <goodwi...@googlemail.com>
>> wrote:
>> > Everyone knows that faceplants are no substitute for pull ups. But I
>> > shouldn't get too excited. I can only do about 6 pull ups.
>>
>> Hm, you can lift 2x more on bench press than I do, but I can do twice
>> as many pullups as you do?
>> How is that possible? I weight 210 lbs.
>
>
> It's just whatever you practice. I used to not be able to do dips. It took
> me months before I could do one. A few weeks later I could do 12.

Ah I see. Yes, I used to be able to only do 3 pullups, but I've been
practicing them every 2-3 days since, so I can do around 12.

Ondrej

[kro...@uni-math.gwdg.de]

 rjf:

 

There are ways of estimating the number of bugs not yet discovered, in case you are interested.

Yes I am

[kro...@uni-math.gwdg.de]

RJF:

[...]

If there is an academic or technical paper that explains how sympy
developers recognize the shortcomings of the design of previous CAS
and how they are overcoming them,

probably there is not and maybe the developers are too young to think about
others experience.
And after they get some wisdom, they are maybe too old to change a lot or to follow the technical progress
except by teaching the new generation.
So somehow things seem still get (very slowly) improved, maybe just by the survival of the fittest?

By the way, the Singular CAS hits the same issue with global flags as Macsyma did, so in that point the Singluar-team
did not learn from the past experience. Greetings to the Singular-team from sage-flame!


Jack

[Bill Hart]

You think this is something that might've been recorded in an academic paper? 

There are probably tens of thousands of bugs in Sage, maxima, singular and sympy. How many of these do you think I should be intimately familiar with before starting my own CAS.

It sounds to me like I will be old before I even start!

Jack

[kro...@uni-math.gwdg.de]

Bill Hart:

[...]

There are probably tens of thousands of bugs in Sage, maxima, singular and
sympy. How many of these do you think I should be intimately familiar with
before starting my own CAS.

It sounds to me like I will be old before I even start!

 

Mea culpa, I didn't study main issues of existing or vanished CAS, did you?
But very likely many of us did study exam protocols or old exam items before attending one. I did.
Is that not mad?

I think it is already helpful to have little background in developing software before starting a project
and then the willingness to learn from others. Nobody can experience everything, so usually we
follow some established behaviour pattern and correct it after some failures.

RJF:

[...]

If there is an academic or technical paper that explains how sympy
developers recognize the shortcomings of the design of previous CAS
and how they are overcoming them,

 

You think this is something that might've been recorded in an academic paper?

 
mhh, you may ask RJF or an internet search engine to get an answer if technical papers about CAS shortcomings exist.

Jack

[rjf]

Here's one
http://www.softwaretestingclub.com/forum/topics/using-the-lincoln-index-as-an

Though like most "metrics" techniques discussed in software engineering, it is
easy to punch holes in it.  Google search will find other methods.

For those too lazy to follow the link: 
Bill finds 12 bugs in a program.
Tom finds 15 bugs.
10 of the bugs are the same.
So we estimate that 12*15/10 , or 18  is the total number of bugs. 

Among other things it assumes that bugs are equally hard to find.

Another issue is that if Bill and Tom find unique bugs -- no overlap,
then the estimate divides by 0.  Infinite number of bugs...

Why is this in sage-flame?  I would have assumed that such
software engineering lore was part and parcel of the background of
builders of large systems intended to be well engineered,
verified etc etc.
RJF

[rjf]

On Saturday, January 18, 2014 2:19:44 AM UTC-8, Bill Hart wrote:

On 18 January 2014 10:55, <kro...@uni-math.gwdg.de> wrote:

RJF:

[...]

If there is an academic or technical paper that explains how sympy
developers recognize the shortcomings of the design of previous CAS
and how they are overcoming them,

probably there is not and maybe the developers are too young to think about
others experience.

It just seems to me that learning about others' attempts to do approximately the same
thing would be a good way of getting ahead faster. 

If we were talking about, say, climbing Mt. Everest, would you expect new young climbers
to ignore existing routes to the summit?  (Not necessarily to follow them, but to know about them.)
 

And after they get some wisdom, they are maybe too old to change a lot or to follow the technical progress
except by teaching the new generation.

Hm.  Did the new generation read the papers describing the old systems, and their problems, at all?
Or did they just plow ahead and think that everything they need to know they learned in high school?
Of course specific algorithms for well-defined problems like polynomial factorization can be copied
out of previous programs or adapted from texts. 
But questions about (say) how should coercion  work... are also important.
What should a database of assumptions look like?

So somehow things seem still get (very slowly) improved, maybe just by the survival of the fittest?

In this situation I would prefer Intelligent Design.  In fact, there is substantial evidence that things do
not necessarily get improved. Furthermore, survival is affected by many issues not related to what
we might think of as fittest  (e.g. mathematically correct).  Was MS-DOS the fittest operating system of the time?

 

By the way, the Singular CAS hits the same issue with global flags as Macsyma did, so in that point the Singluar-team
did not learn from the past experience. Greetings to the Singular-team from sage-flame!

Can you be more specific about this?   Do you think that the builders of Macsyma had no opinions on this?
See
http://www.cs.berkeley.edu/~fateman/papers/mac82b.pdf
 

You think this is something that might've been recorded in an academic paper? 

Sure.

There are probably tens of thousands of bugs in Sage, maxima, singular and sympy. How many of these do you think I should be intimately familiar with before starting my own CAS.

I think you should be familiar with the DESIGN fundamentals of these systems, and especially those
of Mathematica, Maple, MathCAD, Matlab, ... where a supposedly professional team of designers and
programmers tried to design a unified, coherent, useable system.
Although there may be a lot to learn from some of them, academic projects sometimes suffer by becoming a merger of dissertation programs.

I suggest you change your perspective from  thinking of a new CAS as being just like a previous one but without the bugs.

I have so far avoided using Sage directly, but from seeing some bug reports it seems that the
"type" system is --to put it mildly --  confusing.

I have more experience with Mathematica, and have written a review illustrating numerous issues.
For the most part they are not "bugs" and cannot be fixed by "bug fixes".  They can be "removed" by
declaring them to be "features".  But they violate fundamental mathematical tenets, like the
definition of equality.

Are they issues in Sage as well?  Or Singular?  I don't know.

I suspect, but do not know for sure, that some issues like "division by zero" are just tossed up in the air
in Sage.  There are so many subsystems in which one might divide by zero  (Maxima being one).

It sounds to me like I will be old before I even start!

Well, would you object to reading a mathematics book on the grounds that it takes too long, and you
would rather just try to make up your own theorems and proofs?  And that you would be saving time??
 
RJF

Jack

[rjf]

On Saturday, January 18, 2014 6:17:13 AM UTC-8, kro...@uni-math.gwdg.de wrote:

Bill Hart:

[...]

There are probably tens of thousands of bugs in Sage, maxima, singular and
sympy. How many of these do you think I should be intimately familiar with
before starting my own CAS.

It sounds to me like I will be old before I even start!

 

Mea culpa, I didn't study main issues of existing or vanished CAS, did you?

Um, in 1967 there were a few, and yes I studied them.  Formac, Reduce, PM-1, Mathlab, SIN, SAINT,
 Symbolic Mathematical Laboratory, Grad Assistant, Korsvold's simplifier.
 

But very likely many of us did study exam protocols or old exam items before attending one. I did.
Is that not mad?

I think it is already helpful to have little background in developing software before starting a project
and then the willingness to learn from others. Nobody can experience everything, so usually we
follow some established behaviour pattern and correct it after some failures.

By trivializing the initial design process you make the subsequent development more
difficult, and probably more prone to failure.

RJF:

[...]

If there is an academic or technical paper that explains how sympy
developers recognize the shortcomings of the design of previous CAS
and how they are overcoming them,

 

You think this is something that might've been recorded in an academic paper?

 
mhh, you may ask RJF or an internet search engine to get an answer if technical papers about CAS shortcomings exist.

Yes.

Jack

[Tom Boothby]

That's great RJF, I'm really glad you think so. Your contributions
are exemplary as always.

Anyways, I finally got around to doing some pull-ups last night: 16,
followed by 7 after a 1 minute breather.

[Ondřej Čertík]

16 is very good!

Sent from my mobile phone.

[William Stein]

All this talk about pull-ups reminds me of this

http://www.youtube.com/watch?v=loGM3VvpsOE#t=45

[Tom Boothby]

When I was in high school, there was a pull-up bar in the doorway to
the computer room. Thus, I got good at pull-ups and aught else. I
once knocked 25 out in one go... and anything less is a
disappointment. Much like RJF.

On Mon, Jan 20, 2014 at 5:04 PM, Ondřej Čertík <ondrej...@gmail.com> wrote:

[rjf]

My youngest grandchild used pull-ups briefly before she learned to use the
toilet.

That information is also off topic for this group, but perhaps adds a bit of piquancy to this thread.

[Tom Boothby]

Pish posh, toilet humor is always on topic. Piquant? Tell me you
stopped at smelling it...

[Dima Pasechnik]

On 2014-01-17, Bill Hart <goodwi...@googlemail.com> wrote:
> Everyone knows that faceplants are no substitute for pull ups. But I
> shouldn't get too excited. I can only do about 6 pull ups.
>

lately I seem to prefer facepalms...