no one admits bad design decisions

[Ralf Stephan]

For example ignoring that symbolics is the glue
of every CAS; chosing Maxima and not having an
exit strategy; making symbolics a ring when it
is not and then wondering why comparisons with
zero of symbolic matrix elements slow everything
to a crawl; morons who think this is symptomatic
of symbolics and keep on whining about it instead
of using a debugger to find out what and who's
the culprit; everyone who has stopped reaching
for the high-hanging fruits in Sage.

Tucholsky, a free journalist in the 3rd reich,
once said (my translation):

I can't even gorge as much as I want to puke.

[Nathann Cohen]

For example ignoring that symbolics is the glue
of every CAS; chosing Maxima and not having an
exit strategy; making symbolics a ring when it
is not and then wondering why comparisons with
zero of symbolic matrix elements slow everything
to a crawl; morons who think this is symptomatic
of symbolics and keep on whining about it instead
of using a debugger to find out what and who's
the culprit; everyone who has stopped reaching
for the high-hanging fruits in Sage.

I don't know about admitting the mistake in one's own design decisions, but the good thing of not having a boss nor a supreme leader is that we can say exactly that without any problem.

I always thought that some parts of Sage were very hard to develop because "those who need it cannot be expected to know how to write it", i.e.: a lot of people wants to *use* symbolics but have no idea how to write a good system to handle it. Same for some numerical computations.

But yeah, unless we do something about it, symbolics will be the end of Sage.

Nathann

[rjf]

On Saturday, August 1, 2015 at 1:35:03 AM UTC-7, Nathann Cohen wrote:

 

I don't know about admitting the mistake in one's own design decisions, but the good thing of not having a boss nor a supreme leader is that we can say exactly that without any problem.

The effort to maintain a fork of Sage would be substantial, so there

is a  de facto supreme leader. 

I think that the strength and weakness of Mathematica is attributable

to control of design by essentially one person.

As for having an exit strategy for Maxima, I thought sympy was going to be it.

Not that sympy can actually be expected to do better than Maxima overall,

just slightly different.

I always thought that some parts of Sage were very hard to develop because "those who need it cannot be expected to know how to write it", i.e.: a lot of people wants to *use* symbolics but have no idea how to write a good system to handle it. Same for some numerical computations.

Those who think they know how to write it may not have sufficient

understanding of the viewpoint of those who merely need to use it.

 

But yeah, unless we do something about it, symbolics will be the end of Sage.

I'm not sure what this means.  Are you saying that there are Sage contributors who

know how to do symbolics (whatever that includes) right, and they must "do something"?

Did Axiom do it right?

I think that "symbolic" math includes all of computation, with "numeric" as a special

(important) subset.  But also graphics and  (just look at Mathematica)  everything else

from web page hosting to parallel computing to real-time financial databases...

RJF

 

[Tom Boothby]

On Sat, Aug 1, 2015 at 7:48 AM, rjf <fat...@gmail.com> wrote:
>
> The effort to maintain a fork of Sage would be substantial, so there
> is a de facto supreme leader.

Except our particular "supreme leader" is more like a "supreme
rallying point." The only time I've seen William unhappy with people
is when they dick other people over. Never when they disagree with
his choices. He welcomes that sort of debate with open arms, in my
experience.

[William Stein]

I don't think rjf was claiming I'm said de facto supreme leader...?

>
> --
> 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/d/optout.



--
William (http://wstein.org)

[Richard Fateman]

Leader != dictator.
maybe "supreme leader" in my statement should be just "leader"
so as to not invite comparison to the North Korean government.

Presumably 100% complete consensus cannot be the requirement.
(I think one gets around this by referring to the losers as dicks?)

In any case, I think Sage is fairly heavily invested in its own and
its libraries' (sometimes bad) design decisions.

[Tom Boothby]

Majority of informal polls seems to get the job done, when it's deemed
important enough to need more than the nod of a single reviewer. Sage
is very nearly an anarchy. And as far as I can tell, you're the only
person "we" call a dick, where "we" is me, and only because it's a
common nickname for Richards.

So yes, I see and agree with your point (let it be known that this
happens): like every other large project, we are ruled more by
momentum than by any individual.

In response to the title of this thread... bullshit. There's an
active "mea culpa" thread going on right now, providing a direct
contradiction to the claim in the title. Moreover, an important part
of any judgement about a past decision is an evaluation of historical
context. Picking maxima *was* the best choice at the time. The
results of several of the corrollary decisions kinda suck (not that
they're necessarily wrong, just that they have bad side-effects), but
a dedicated user can learn how to avoid maxima except when absolutely
necessary.

[Richard]

On 8/4/15, 11:41 AM, Tom Boothby wrote:
> Majority of informal polls seems to get the job done, when it's deemed
> important enough to need more than the nod of a single reviewer. Sage
> is very nearly an anarchy. And as far as I can tell, you're the only
> person "we" call a dick, where "we" is me, and only because it's a
> common nickname for Richards.

No problem, I can decide to call you (a) dick.

So some (leader?) runs a poll. Anyone can vote? A majority of people
vote in favor of demonstrably factual false positions. Determining that
a design decision is good, or that a previous design decision is bad ...
maybe not something best done by a poll or an informal poll (what is
that -- ask your office mates and see if they agree?)

>
> So yes, I see and agree with your point (let it be known that this
> happens): like every other large project, we are ruled more by
> momentum than by any individual.

Dunno about "every other large project".

>
> In response to the title of this thread... bullshit. There's an
> active "mea culpa" thread going on right now, providing a direct
> contradiction to the claim in the title. Moreover, an important part
> of any judgement about a past decision is an evaluation of historical
> context.

Not a thread I see. Unless you mean this very one on sage-flame.

> Picking maxima *was* the best choice at the time.

This is not at the level of design decision I was thinking about. There are
so many. Like do you assume computation is done over the reals, complex
numbers,
etc. Or how do you terminate a command. Or do you use TeX or LaTeX.
Or how do you use "assumptions".
Or what to do with numeric overflows.

Using Maxima -- well maybe you could use Axiom.
Given the legal licensing requirements for pieces of Sage, which I view as
a major design decision, who knows. It would be plausible to list pro/con
but frankly, I think that decision, even if, in retrospect, is doubtful,
would
not be changed.

Rjf

[William Stein]

On Tue, Aug 4, 2015 at 12:21 PM, Richard <fat...@gmail.com> wrote:
> On 8/4/15, 11:41 AM, Tom Boothby wrote:
>>
>> Majority of informal polls seems to get the job done, when it's deemed
>> important enough to need more than the nod of a single reviewer. Sage
>> is very nearly an anarchy. And as far as I can tell, you're the only
>> person "we" call a dick, where "we" is me, and only because it's a
>> common nickname for Richards.
>
> No problem, I can decide to call you (a) dick.
>
> So some (leader?) runs a poll. Anyone can vote? A majority of people
> vote in favor of demonstrably factual false positions. Determining that
> a design decision is good, or that a previous design decision is bad ...
> maybe not something best done by a poll or an informal poll (what is
> that -- ask your office mates and see if they agree?)
>

When there are contentious issues (e.g., the code of conduct), we have
a majority vote.

>> In response to the title of this thread... bullshit. There's an
>> active "mea culpa" thread going on right now, providing a direct
>> contradiction to the claim in the title. Moreover, an important part
>> of any judgement about a past decision is an evaluation of historical
>> context.
>
> Not a thread I see. Unless you mean this very one on sage-flame.
>

He might be alluding to me saying that our packaging system is archaic?

> Given the legal licensing requirements for pieces of Sage, which I view as
> a major design decision, who knows.

The license of sage is a design decision, but not one made by Sage
developers. It was made by people like Henri Cohen, who licensed PARI
under the GPL (or somebody like you and Maxima???), which forces Sage
to be licensed under the GPL. We had no say in the matter... It was
the previous generation's.


--
William (http://wstein.org)