Sources of funding - perhaps computer manufacturers?

[Dr. David Kirkby (Kirkby Microwave Ltd)]

A ticket I opened years ago about a problem on AIX, got closed recently. It got me thinking about something  whose usefulness could well extend beyond one issue building Sage on AIX.

I gather William is having problems getting funding from NSF and similar places. I wonder if it's time to look at this a different way. Based on things in the past

1) Sun sponsored a port to Solaris, and paid the salary of someone. Unfortunately the computer they donated (t2) was not suited to the task, but that is irrelevant now. We did eventually get Sage ported to Solaris.

2) Someone from IBM contacted William some time ago an IBM funded port to AIX. I got involved, as I did have an AIX box, but nothing ever came of it.

I am realistic, and don't expect many Sage developers to care less about AIX, although I think there is at least one other that will do. IBM do have some nice hardware.

But how about contacting manufacturers of other devices, to sponsor either a full Sage port, or a subset of Sage.

Some that come to mind are

i) Nokia

ii) Samsung

iii) Apple

iv) Microsoft

v) Oracle, with their own flavor of Linux.

vi) Cray - obviously concentrating on parallel processing

Then there's the possibility of a mobiles apps for Android and Apple phones that have a subset of functionality without internet access, and better access with internet access.

Dr. David Kirkby Ph.D CEng MIET
Kirkby Microwave Ltd
Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Essex, CM3 6DT, UK.
Registered in England and Wales, company number 08914892.
http://www.kirkbymicrowave.co.uk/
Tel: 07910 441670 / +44 7910 441670 (0900 to 2100 GMT only please)

[William Stein]

On Sunday, September 13, 2015, Dr. David Kirkby (Kirkby Microwave Ltd) <drki...@kirkbymicrowave.co.uk> wrote:

A ticket I opened years ago about a problem on AIX, got closed recently. It got me thinking about something  whose usefulness could well extend beyond one issue building Sage on AIX.

I gather William is having problems getting funding from NSF and similar places. I wonder if it's time to look at this a different way. Based on things in the past

1) Sun sponsored a port to Solaris, and paid the salary of someone. 

You are misremembering slightly.  Sun didn't give us a penny and definitely didn't fund Michael Abshoff to work on the port.  Sun gave UW one computer and some publicity, then a few months later they got bought buy oracle and all education outreach went silent.  

 

Unfortunately the computer they donated (t2) was not suited to the task, but that is irrelevant now. We did eventually get Sage ported to Solaris.

2) Someone from IBM contacted William some time ago an IBM funded port to AIX. I got involved, as I did have an AIX box, but nothing ever came of it. 

We exchanged emails back and forth for a while which got my hopes up temporarily but it was very clear there would be $0 support there.   

 

I am realistic, and don't expect many Sage developers to care less about AIX, although I think there is at least one other that will do. IBM do have some nice hardware.

But how about contacting manufacturers of other devices, to sponsor either a full Sage port, or a subset of Sage.

Some that come to mind are

i) Nokia

ii) Samsung

iii) Apple

iv) Microsoft

v) Oracle, with their own flavor of Linux.

vi) Cray - obviously concentrating on parallel processing

Then there's the possibility of a mobiles apps for Android and Apple phones that have a subset of functionality without internet access, and better access with internet access.

Dr. David Kirkby Ph.D CEng MIET
Kirkby Microwave Ltd
Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, Essex, CM3 6DT, UK.
Registered in England and Wales, company number 08914892.
http://www.kirkbymicrowave.co.uk/
Tel: 07910 441670 / +44 7910 441670 (0900 to 2100 GMT only please)

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To post to this group, send email to sage-...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

--
Sent from my massive iPhone 6 plus.

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 13 September 2015 at 18:31, William Stein <wst...@gmail.com> wrote:

On Sunday, September 13, 2015, Dr. David Kirkby (Kirkby Microwave Ltd) <drki...@kirkbymicrowave.co.uk> wrote:

A ticket I opened years ago about a problem on AIX, got closed recently. It got me thinking about something  whose usefulness could well extend beyond one issue building Sage on AIX.

I gather William is having problems getting funding from NSF and similar places. I wonder if it's time to look at this a different way. Based on things in the past

1) Sun sponsored a port to Solaris, and paid the salary of someone. 

You are misremembering slightly.  Sun didn't give us a penny and definitely didn't fund Michael Abshoff to work on the port.  Sun gave UW one computer and some publicity, then a few months later they got bought buy oracle and all education outreach went silent.  

OK, I was mistaken. It is a long time ago - back in 2009 we got t2.math I think.

 

 

Unfortunately the computer they donated (t2) was not suited to the task, but that is irrelevant now. We did eventually get Sage ported to Solaris.

2) Someone from IBM contacted William some time ago an IBM funded port to AIX. I got involved, as I did have an AIX box, but nothing ever came of it. 

We exchanged emails back and forth for a while which got my hopes up temporarily but it was very clear there would be $0 support there.   

Agreed, the AIX interested from someone at IBM did go anywhere. But that does not mean that an approach to other hardware/software vendors would fail. If funding has dried up from research grants, perhaps another approach is needed.

Dave

[William Stein]

I won't be pursuing this.   I can only do so many things at once, and not focusing would ensure failure.  If somebody else wants to try, please go for it! 

 

Dave

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To post to this group, send email to sage-...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 13 Sep 2015 21:49, "William Stein" <wst...@gmail.com> wrote:

>> Agreed, the AIX interested from someone at IBM did go anywhere. But that does not mean that an approach to other hardware/software vendors would fail. If funding has dried up from research grants, perhaps another approach is needed.
> I won't be pursuing this.   I can only do so many things at once, and not focusing would ensure failure.  If somebody else wants to try, please go for it! 

I can't as I don't work in a uni, but perhaps others could.  Anyway it was just an idea.

Dave

[rjf]

Unless you can argue that having a Sage port will increase sales, then the marketing types

won't care. 

Unless you can argue that giving money to a university is a better way to pursue

a research topic of interest, then the research types would rather pay in-house.

If they were at all interested.

The much-maligned controversial statement (Sage doomed) worth reviewing?

http://sci.tech-archive.net/Archive/sci.math.symbolic/2005-12/msg00096.html

A quote..

By avoiding applications (say, to engineering design, finance,
education, scientific visualization, etc etc) the activity is
essentially doomed. Why? Government funding for people or projects
will be a small percentage of the funding for pure mathematics.
That's not much. And the future is pretty grim.

..............

There is more money now than 10 years ago in the hands of 

mathematically-friendly philanthropists.  Simons, Clay, Beale, maybe

Paulson.

What would motivate them?

RJF

[William Stein]

On Sunday, September 27, 2015, rjf <fat...@gmail.com> wrote:

Unless you can argue that having a Sage port will increase sales, then the marketing types

won't care. 

Unless you can argue that giving money to a university is a better way to pursue

a research topic of interest, then the research types would rather pay in-house.

If they were at all interested.

The much-maligned controversial statement (Sage doomed) worth reviewing?

Though I think the growth/survival of the SageMath project is in serious jeopardy due to lack of funding, I for one am definitely not giving up.  The activity is not doomed.  If anything the only thing to do is try much harder, try a wider range of approaches to getting support, be more open minded, etc. Also the recent success of many sage devs in getting the $8m OpenDreamKit grant is very inspiring!  As are the 161 paying customers of SageMathCloud.  And the innovations like Bill Hart's Nemo project.   Long live open source mah software, starting long ago with Maxima :-)

William 

 

http://sci.tech-archive.net/Archive/sci.math.symbolic/2005-12/msg00096.html

A quote..

By avoiding applications (say, to engineering design, finance,
education, scientific visualization, etc etc) the activity is
essentially doomed. Why? Government funding for people or projects
will be a small percentage of the funding for pure mathematics.
That's not much. And the future is pretty grim.

..............

There is more money now than 10 years ago in the hands of 

mathematically-friendly philanthropists.  Simons, Clay, Beale, maybe

Paulson.

What would motivate them?

RJF

--

You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To post to this group, send email to sage-...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[Nathann Cohen]

> Though I think the growth/survival of the SageMath project is in serious
> jeopardy due to lack of funding, I for one am definitely not giving up. The
> activity is not doomed. If anything the only thing to do is try much
> harder, try a wider range of approaches to getting support, be more open
> minded, etc. Also the recent success of many sage devs in getting the $8m
> OpenDreamKit grant is very inspiring! As are the 161 paying customers of
> SageMathCloud. And the innovations like Bill Hart's Nemo project. Long
> live open source mah software, starting long ago with Maxima :-)

Do not blame me for that, but to me (*) Sage is before anything else a
tool for researchers. To experiment, to compare, to check things. And
on that front it seems to have been a while since new folders were
created in src/. The most recent I could name are the finite automata
guys, and perhaps the asymptotic expressions thing. Is the domain of
mathematics covered by Sage expanding these days? What about the
manifold guys, for instance? Are they joining the project or do they
still develop on their own?

This lack of mathematical expansion bothers me.

Nathann

(*) It is very important to read "to me" and nothing else. I make no
claim on what Sage *should* be, or what people *should* believe it to
be.

[William Stein]

On Sunday, September 27, 2015, Nathann Cohen <nathan...@gmail.com> wrote:

> Though I think the growth/survival of the SageMath project is in serious
> jeopardy due to lack of funding, I for one am definitely not giving up.  The
> activity is not doomed.  If anything the only thing to do is try much
> harder, try a wider range of approaches to getting support, be more open
> minded, etc. Also the recent success of many sage devs in getting the $8m
> OpenDreamKit grant is very inspiring!  As are the 161 paying customers of
> SageMathCloud.  And the innovations like Bill Hart's Nemo project.   Long
> live open source mah software, starting long ago with Maxima :-)

Do not blame me for that, but to me (*) Sage is before anything else a
tool for researchers. To experiment, to compare, to check things. And
on that front it seems to have been a while since new folders were
created in src/. The most recent I could name are the finite automata
guys, and perhaps the asymptotic expressions thing. Is the domain of
mathematics covered by Sage expanding these days? What about the
manifold guys, for instance? Are they joining the project or do they
still develop on their own?

This lack of mathematical expansion bothers me.

+1

Thanks for pointing this out as a way for Sage to improve. 

I think sage-manifolds is being developed very much right now, but isn't included in sage.   I personally think sage (like Python) would greatly benefit by having a healthy third party package infrastructure, with lots of code that is NOT necessarily included in sage, but is of high quality and easy to install.   Many (most?!) other software ecosystems have thrived without having to be monolithic.  

 

Nathann

(*) It is very important to read "to me" and nothing else. I make no
claim on what Sage *should* be, or what people *should* believe it to
be.

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To post to this group, send email to sage-...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

[Eric Gourgoulhon]

Le dimanche 27 septembre 2015 20:20:18 UTC+2, William a écrit :

On Sunday, September 27, 2015, Nathann Cohen <nathan...@gmail.com> wrote:

 What about the
manifold guys, for instance? Are they joining the project or do they
still develop on their own?

I think sage-manifolds is being developed very much right now, but isn't included in sage. 

... but should be at some point: have a look at
http://trac.sagemath.org/ticket/18528
This bunch of tickets, which introduces the new folder src/sage/manifolds ;-), is not ready
for review yet (a few improvements in the documentation are under way), but should
be very soon. I'll then post a message on sage-devel.
Let me mention that from the very beginning of the SageManifolds project, the aim was to
have it fully included in Sage. In particular, everything follows the parent/element scheme.
Note by the way that the pure algebraic part (tensors on free modules) is already included
in Sage (since version 6.6).

Eric.

[kcrisman]

The much-maligned controversial statement (Sage doomed) worth reviewing?

Though I think the growth/survival of the SageMath project is in serious jeopardy due to lack of funding, I for one am definitely not giving up.  The activity is not doomed. 

As usual, I think it's important to distinguish between the survival of Sage - I don't think it's in any more danger than Maxima or many other projects of that. Growth beyond its current user base, that is the more tricky part.  

[William Stein]

Exactly.  And also the mission statement: viable alternative to the Ma's - that is tricky!

[Francesco Biscani]

Exactly.  And also the mission statement: viable alternative to the Ma's - that is tricky!

I have always felt a tad confused and mislead by this statement.

As someone who has interacted over the years with physicists and engineers using daily Mathematica, Maple and Matlab, I see very little overlap between their typical use of these tools and the typical usages of SAGE, at least from the point of view of a lurker on this list. It seems like SAGE caters to (and is run mostly by) researchers in pure mathematics, and that is little interest on other use cases. Pragmatically, it seems to me that a sizeable chunk of people "doing mathematics on a computer" is today better served in the Python space by the Numpy/SciPy/SymPy/Matplotlib stack as an alternative to the Ma's rather than SAGE.

This is of course completely fine! I am not questioning anyone's motives, inclinations or desires. But IMO continuing to push the idea that SAGE aims to be a viable alternative to the Ma's tout-court risks of being a source of confusion.

Cheers,

  Francesco.

[William Stein]

1. Magma is also an Ma. Magma's incredibly good at pure mathematics.
You seem to be leaving out Magma above.

2. You say "... better served in the Python space by the

Numpy/SciPy/SymPy/Matplotlib stack as an alternative to the Ma's

rather than SAGE." Sage includes "Numpy/SciPy/SymPy/Matplotlib", so
we don't have to worry about that chunk of people with respect to our
mission statement.

3. There is a lot more to mathematics than just what Magma does and
*also* much more to it than just what Numpy/SciPy/SymPy/Matplotlib do.
There's a huge amount of interesting things that could be
systematically computed with in mathematics that no existing package
does yet.

-- William

>
> Cheers,
>
> Francesco.

>
>
> --
> You received this message because you are subscribed to the Google Groups
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sage-devel+...@googlegroups.com.
> To post to this group, send email to sage-...@googlegroups.com.
> Visit this group at http://groups.google.com/group/sage-devel.
> For more options, visit https://groups.google.com/d/optout.



--

William (http://wstein.org)

[Francesco Biscani]

On 28 September 2015 at 19:37, William Stein <wst...@gmail.com> wrote:

1. Magma is also an Ma.   Magma's incredibly good at pure mathematics.
You seem to be leaving out Magma above.

I admit I know basically nothing about Magma (I did not know it even existed before joining this list :).

 

2. You say "... better served in the Python space by the
Numpy/SciPy/SymPy/Matplotlib stack as an alternative to the Ma's
rather than SAGE."   Sage includes "Numpy/SciPy/SymPy/Matplotlib", so
we don't have to worry about that chunk of people with respect to our
mission statement.

I can install the NSPM stack on any modern platform (Windows included) with pip (or my distro's package manger, if I am on linux) in probably less than 5 minutes with a reasonable internet connection. I don't have to worry about sizeable downloads, virtual machines, containers, emulation layers a-la cygwin, installing a separate compiler toolchain/python version/set of libraries, or anything of the sort.

It is true that you can use the NSPM stack from SAGE, but what are the key advantages of doing so? It is a honest question, maybe there's something I am overlooking.

3. There is a lot more to mathematics than just what Magma does and
*also* much more to it than just what Numpy/SciPy/SymPy/Matplotlib do.
  There's a huge amount of interesting things that could be
systematically computed with in mathematics that no existing package
does yet.

Yes, nobody has the monopoly on what "mathematics on a computer" means :)

My comment was merely a marketing/strategic one: I think there exists a disconnect between the mission statement and what SAGE actually is (and maybe what the SAGE community wants it to be).

Cheers,

  Francesco.

[Bill Page]

Given the serious situation in Sage funding I suppose that there is
still a good reason for continuing this thread.

On 28 September 2015 at 13:37, William Stein <wst...@gmail.com> wrote:
>
> On Mon, Sep 28, 2015 at 10:25 AM, Francesco Biscani
> <blues...@gmail.com> wrote:
> >> Exactly. And also the mission statement: viable alternative to the Ma's -
> >> that is tricky!
> >
> > I have always felt a tad confused and mislead by this statement.
> >
> > As someone who has interacted over the years with physicists and engineers
> > using daily Mathematica, Maple and Matlab, I see very little overlap between
> > their typical use of these tools and the typical usages of SAGE, at least
> > from the point of view of a lurker on this list.

> ...

> 1. Magma is also an Ma. Magma's incredibly good at pure mathematics.
> You seem to be leaving out Magma above.
>

With emphasis on "physicists and engineers" I completely agree with
Francesco. I am not aware of any physicists or engineers who use
Magma. I never heard of Magma before Sage and I still find Magma of
little interest - for physics or engineering. Perhaps I just do not
know what I am are missing? Meanwhile I admit that I do know something
about Axiom, another system that might be accused of catering to pure
mathematics, and I have used it in theoretical physics. And I have
also used Sage, or more specifically "Numpy/SciPy/SymPy/Matplotlib"
(not to forget also Maxima and probably several other packages
transparently wrapped up in Sage) on SMC.

Because my collaborator has less tolerance for the current
idiosyncrasies of SMC and Sage and a greater familiarity with Maple, I
recently back ported one of my more complicated Sage worksheets to
Maple. I found it a bit challenging. I have also used Maple for a
long time but it turned out that I had used some features in Sage and
Numpy for which I did not immediately know the Maple counterpart.
However the final result was just fine and convinced me that in many
ways Sage is definitely an alternative to Maple even though it may
seem more viable to some people than others.

> > It seems like SAGE caters
> > to (and is run mostly by) researchers in pure mathematics, and that is
> > little interest on other use cases. Pragmatically, it seems to me that a
> > sizeable chunk of people "doing mathematics on a computer" is today
> > better served in the Python space by the Numpy/SciPy/SymPy/Matplotlib
> > stack as an alternative to the Ma's rather than SAGE.
> >

> 2. You say "... better served in the Python space by the
> Numpy/SciPy/SymPy/Matplotlib stack as an alternative to the Ma's
> rather than SAGE." Sage includes "Numpy/SciPy/SymPy/Matplotlib",
> so we don't have to worry about that chunk of people with respect to
> our mission statement.
>

This seems odd from the point of view of marketing strategy. If
"Numpy/SciPy/SymPy/Matplotlib" is already an alternative to the Ma's
(minus Magma), then is the only point of Sage to add the missing
features of Magma? In terms of attracting funding for Sage, I would
be worried about showing that Sage provides some obvious added value
over just "Numpy/SciPy/SymPy/Matplotlib" for those users.

In this regard it is kind of interesting to read:

https://github.com/sympy/sympy/wiki/SymPy-vs.-Magma

Of course it is a kind of advertisement for Sympy, but something like
this might be appropriate for Sage.

SMC as a platform on the other hand seems much more agnostic and
hopefully is attract some of these users, although there does seem to
be some significant competition using a similar tool set.

> > This is of course completely fine! I am not questioning anyone's motives,
> > inclinations or desires. But IMO continuing to push the idea that SAGE
> > aims to be a viable alternative to the Ma's tout-court risks of being a
> > source of confusion.
>

> 3. There is a lot more to mathematics than just what Magma does and
> *also* much more to it than just what Numpy/SciPy/SymPy/Matplotlib
> do. There's a huge amount of interesting things that could be
> systematically computed with in mathematics that no existing package
> does yet.
>

The point being that this is not explicitly part of the Sage "mission
statement". Of course there are quite a few people who seem to be
trying to do this with Sage but I am not sure whether Sage is more or
less an viable alternative for this purpose than an of the Ma's. When
it comes to doing new mathematics the flexibility of Python and the
complexity of the Sage development infrastructure both seem daunting
compared to the tightly integrated mathematics library in a system
like Axiom (FriCAS).

Bill Page.

[Dima Pasechnik]

On Tuesday, 29 September 2015 10:06:56 UTC-7, Bill Page wrote:

Given the serious situation in Sage funding I suppose that there is
still a good reason for continuing this thread.

On 28 September 2015 at 13:37, William Stein <wst...@gmail.com> wrote:
>
> On Mon, Sep 28, 2015 at 10:25 AM, Francesco Biscani
> <blues...@gmail.com> wrote:
> >> Exactly.  And also the mission statement: viable alternative to the Ma's -
> >> that is tricky!
> >
> > I have always felt a tad confused and mislead by this statement.
> >
> > As someone who has interacted over the years with physicists and engineers
> > using daily Mathematica, Maple and Matlab, I see very little overlap between
> > their typical use of these tools and the typical usages of SAGE, at least
> > from the point of view of a lurker on this list.
> ...
> 1. Magma is also an Ma.   Magma's incredibly good at pure mathematics.
> You seem to be leaving out Magma above.
>

With emphasis on "physicists and engineers" I completely agree with
Francesco. I am not aware of any physicists or engineers who use
Magma.

 

cryptographers (some of them can certainly qualify as engineers) use Magma a lot.

[Bill Page]

On 29 September 2015 at 19:41, Dima Pasechnik <dim...@gmail.com> wrote:
>
> cryptographers (some of them can certainly qualify as engineers) use Magma a lot.
>

OK. Would you say that Sage is a viable alternative for them?

[Ralf Stephan]

On Tuesday, September 29, 2015 at 4:44:07 PM UTC+2, bluescarni wrote:

It is true that you can use the NSPM stack from SAGE, but what are the key advantages of doing so? It is a honest question, maybe there's something I am overlooking.

In calculus you still want to have Maxima's integrator even if symbolics

are nicer with SymPy.  

[William Stein]

No.   Going into details would be too technical but overall sage is not a viable alternative to magma for most crypto researchers.   

 

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To post to this group, send email to sage-...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

--

[Dr. David Kirkby (Kirkby Microwave Ltd)]

I tend to agree,  even though I am aware Sage included  Numpy/SciPy/SymPy/Matplotlib.

Octave, which is a MATLAB clone,  is supported as an optional package, though personally if I wanted to use Octave,  I would run Octave.

One thing that Sage could do with,  which might attract some engineers, is GPIB and RS-232 support to control instruments, which is a small part of the MATLAB instrument control toolbox.

http://uk.mathworks.com/products/instrument/

I think the ability to control instruments from MATLAB, is essential. GPIB (IEEE-488)

https://en.wikipedia.org/wiki/IEEE-488

is one way to communicate with them, and is the only method used on older instruments. Some low-cost modern instruments use USB and/or LAN, but the more expensive modern instruments will use USB, LAN and GPIB. Some instruments can be controlled via RS-232.

I believe a Python plugin exists for GPIB.  A lot of instruments have GPIB connectivity.  If I look in my lab, I have some instruments with GPIB and some without. The list with GPIB support is much longer than without. 

*WITH* GPIB

* HP 8753ES 300 kHz to 3 GHz vector network analyser
* HP 8720D 50 MHz to 20 GHz vector network analyser
* 22 GHz Spectrum analyzer based on HP 70000 modular  measurement system
* 4 x Agilent power supplies, of various voltages & currents
* HP 8970A noise figure meter.
* 20 GHz HP 83623A sweep generator
* 4.2 GHz HP 8665A signal generator
* 1 GHz Marconi 2022D signal generator
* 20 GHz IFR 2187 programmable stepped attenuator.
* Stanford Research DS345 30 MHz function generator
* HP 438A Dual channel power meter.
* EG&G 7260 lock in amplifier.
* HP 4284A precision frequency reference
* 18 GHz HP frequency counter
* HP 3457A 6.5 digit bench multimeter

WITHOUT GPIB
* HP 58503A GPS locked frequency reference - that has RS232 control.
* 2 x handheld Tektronix 4.5 digit multimeters.
* Peak ESR70 meter.
* HP 100 MHz HP oscilloscope. - although GPIB is an option on this. 

I guess adding the GPIB module into Sage, would allow instruments to be controlled, would be one small step towards making it attractive to engineers, especially if there was an online demo of real time streaming of data from some test equipment. But realistically, Francesco is right, there's not a lot to attract engineers using MATLAB or Mathematica.

Personally my experience is most engineers use MATLAB or Labview - I see very few using Mathematica or Maple. 

It's a bit of a chicken-and-egg situation. Until you get more engineers using Sage, you wont get the engineering tools engineers need. Until those tools exists, you wont get engineers using Sage. 

So Francesco is right. With Sage not having anywhere near the functionality of MATLAB for engineers, it is not going to attract them, so it is not a viable alternative. Just look at the toolboxes available for MATLAB

http://uk.mathworks.com/products/

any you will soon see Sage is far from a viable alternative to MATLAB. Sage may be more of a viable alternative to Mathematica and Maple.

Dr David Kirkby
(A chartered  engineer but *not* a mathematician)

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 30 September 2015 at 10:07, Dr. David Kirkby (Kirkby Microwave Ltd) <drki...@kirkbymicrowave.co.uk> wrote:

O * HP 4284A precision frequency reference

Oops, the 4284A is a precision LCR meter, not a frequency reference. That instrument is obsolete

http://www.keysight.com/en/pd-1000000874:epsg:pro-pn-4284A/precision-lcr-meter-20-hz-to-1-mhz

but with a basic uncertainly of 0.05%, has much lower uncertainty than many more modern instruments. If you exclude cheap Chinese instruments, where rand() is used to generate the specifications, then there's nothing to touch the 4284A, unless one spends serious amounts of money.

A quick Google will find drivers existing to control the 4284A with MATLAB

http://forums.ni.com/t5/LabVIEW/hp-4284A-vi-s-for-old-Matlab-version-6-1/td-p/871386

I've not bothered checking, but I doubt there would be with Mathematica.

Dave

[Bill Hart]

Money is actually not always the problem.

Here in Europe we currently have multiple positions advertised, including one for developing MPIR. And most of these positions are software engineer positions that will ultimately improve some technology used by Sage (or which could be used by Sage).

MPIR doesn't have really good support for processors later than about 2010 and almost nothing at all post 2012 (after Jason Moxham died). There are unconfirmed reports of libraries being developed that are up to 6 (SIX) times faster than MPIR (and probably GMP too) on modern processors.

We have the money. We know what needs to be done. But we have zero applicants. There is a lack of talent, not a lack of money in some areas.

The fact remains that the Sage community as a whole does not have the resources (here I mainly mean people, not money) to support some of its essential core activities.

I think Sage has diversified too much and has not been able to solve many core problems:

1) How to continue maintaining dependencies after people die/move on, etc.

2) How to keep being a viable alternative after it becomes one. What analysis has been done of the weak spots of Sage? What things can't Sage do without (other than money)? What plans are in place to safeguard those investments?

3) There are many areas in which Sage has not made a dent in the dominance of Magma and similarly for Maple (I know relatively less about the other MA's). Some of these areas are highly mathematical, others are purely technological.

Whilst more money would likely help in some areas. In other areas it won't make the slightest difference. Magma has very few resources currently, but is still steaming ahead (albeit into an uncertain future).

The Sage project simply isn't attracting the right talent in some areas (in others it is, obviously). In some areas, our competition has all the talent. There just isn't any left to buy up. The only thing that could fix this would be training.

RJF is absolutely right with his comment about marketing types. Asking for money from companies is not going to solve the problem. If we don't have any users with that kind of hardware, that is not likely a result of those companies not giving us money, nor will it likely be fixed by them doing so. They will simply see Sage as irrelevant to their marketing.

In my personal opinion, Sage absolutely doesn't want to be diversifying further into dead/outdated platforms, or onto platforms where there has been historically very little expertise available for Open Source development. In lean times like this, we need to focus on safeguarding core activities and on solving core problems.

I think the right question to be asking is this. What does Sage need the most right now?

1) More serious well-known mathematicians using and advocating Sage.

2) Making a dent in the dominance of competitors.

3) Support for more platforms.

4) Technological advances.

5) Development of algorithms.

6) Performance.

7) Tools and ecosystem improvements.

8) Other.

I'm sure opinion will be divided, but I'd put 3 dead last. Of course native Windows support would still be nice, but that's not something that can seriously be considered in lean times like this. And I question whether it would add a single extra skilled developer to the project.

Question: how can Sage hope to remain relevant if Magma and Maple are working on technological improvements that will put Sage a further factor of 6 behind in core areas, before we even start thinking about their multicore and GPU investments?

 

Question: what is Sage's strategy going forward in light of the fact that it is not making a dent in the dominance of other platforms? I mean other than getting more money?

No one is going to give money to a project that doesn't have a worked out strategy (other than getting more money), and especially not if it isn't part of their own strategy!

Bill.

[Bill Hart]

I mean, people understand, right, that in order to make headway in a given area one needs to become the platform of choice for people in that area.

Just being a viable alternative isn't really enough. It's necessary to actually disrupt (that's the relevant buzzword) to a sufficient extent the dominance of the current technology leader.

When it becomes clear that Sage is the platform of choice for that area then the number of users goes up in that area and more developers move to that platform to support their users.

I recently sat down with some serious developers and we discussed symbolics in Sage (which I know nothing about). They argued that Sage is not a viable contender in that area, and we discussed some of the possible reasons for that. The thing is, there is nothing even on the horizon which looks likely to disrupt the technology leaders in that area.

The same can be said about the areas Magma is currently dominant. The best chance Sage has in that direction is if Magma dies due to some stupidity at the University of Sydney. There's nothing I can see technologically in Sage that even has the potential to disrupt Magma's dominance. Not even in theory.

People will point to William's amazing cloud.sagemath.com, but in the final analysis, people will just figure out how to use it to run Magma or Maple. That's great if it means more funding for Open Source software development. But it isn't making real headway with the core problem, which is how to become the platform of choice for serious mathematicians in the core areas of competence of those other projects.

Bill.

[Jori Mäntysalo]

On Wed, 30 Sep 2015, Bill Hart wrote:

> When it becomes clear that Sage is the platform of choice for that area
> then the number of users goes up in that area and more developers move
> to that platform to support their users.

It is usually good to build on strong points. So, what are best areas in
Sage? Where it now is The Software(tm) to use?

And how could we expand those to some near area?

--
Jori Mäntysalo

[Bill Hart]

I think that is a good question. I don't personally know the answers, since my experience is quite limited to the areas that interest me.

However, it is my understanding that a significant number of people moved to Sage to start the sage-combinat project from MuPad. I don't know enough about it to say if that is one of the great success stories. But it is the example I had in mind when I wrote my comments.

One thing Sage doesn't need to fix is BLAS. This area has so much effort poured into it by numerical people (who have deep pockets), independently of Sage, that something like OpenBLAS (or whatever Sage is using instead), probably couldn't be improved by the Sage community even if we tried. The code in that thing is scary fast, supports right up to modern processors, supports threads. It's a technological marvel. So that's a strength for Sage. Nobody will come along and disrupt that.

Flint was for a while a big success for Sage. But it's fallen woefully behind on modern architectures (SIMD) and has very little threading, though it is generally threadsafe.

Arb is definitely a major innovation. However, I think it really exists in an area where there is no competition. This is good for Sage, but not the sort of thing I was talking about.

The L-functions and modular forms database is relying on a lot of Sage technology, and I would say that is good for Sage (though realistically quite a bit of the data there has been computed with Magma, or could be more efficiently computed with Magma). There are some serious mathematicians involved with Sage in that direction.

Bill.

[Nathann Cohen]

To Bill:

Reading your post, it seems that what you consider to be Sage's
strength is not Sage's own code but rather the fact that we make many
different softwares coexist in here.

Seems to give even more reasons to provide easy ways for everybody to
plug things in here.

Perhaps we should rely much more on system-available tools, instead of
requiring Sage packages?

Nathann

[Bill Hart]

I definitely think Sage is all about supporting Open Source software development, rather than a single monolithic project called Sage. That's definitely understandable, since everyone who comes to an Open Source project comes with their own vision of what they'd like to achieve with the project.

Sage is more of a platform than a piece of software. Or at least, I am beginning to see it more that way.

In my opinion, Sage would have done better if it were more modular. I commonly hear the complaint that Sage is "not a Python package", but "you have to get Sage" to run it.

On the other hand, I definitely think Sage tried to diversify too much. It became a Python distribution, without being a Python distribution. A lot of the things Sage has tried to do could perhaps have been better supported as independent Python projects which can be used independently of Sage. (Some are, I know.)

But I have to declare my strong bias here. I would personally have liked to see Sage disrupt Magma significantly and focus on being a viable replacement for Magma. I would have like to see Sage adopted by the majority of serious mathematicians in the areas Magma covers.

As a piece of software, rather than as a distribution, I would have liked to see Sage more focused on that goal. Instead I think it's become too broad. It's 800 mb of software, most of which I don't need on my PC (it doesn't run on there anyway, so no problem).

But what I would have liked to see isn't what the majority of the Sage community envisions when they see Sage. So you are right, Sage is instead a platform which enables people to achieve the specific goals they have in mind, not a single monolithic project which is focused on a certain goal. So a goal of Sage could be to enable more people to put their stuff into the project. But will they?

Bill.

[Volker Braun]

On Wednesday, September 30, 2015 at 4:55:29 PM UTC+2, Bill Hart wrote:

We have the money. We know what needs to be done. But we have zero applicants. There is a lack of talent, not a lack of money in some areas.

You are looking for an expert in compilers / optimization who happens to have a math graduate degree. Those obviously exist, the problem is that that this is a valuable skillset. I'm guessing to the tune of $200k annual if you work for one of the tech giants. Since you can't pay with stock options that means either

a) offer market rate (and 200k for one year only is less attractive than 200k every year)

b) offer flexibility (remote work etc.)

c) offer a chance at tenure (long postdoc, prestigious institution, famous adviser)

Its basic economics: If you didn't get any applications then your compensation is inadequate.

[William Stein]

On Wed, Sep 30, 2015 at 7:55 AM, Bill Hart <goodwi...@googlemail.com> wrote:
> What analysis has been done of the weak spots of Sage?

[...]
> Question: what is Sage's strategy[...]?

Here is Sage's strategy, or at least what my strategy toward Sage has
been for the last 5 years.

1. Diagnose the problem:

Statement of problem: SageMath is not growing.

Justification:

Facts: Growth in the number of active users [1] of SageMath has
stalled since about 2011 (as defined by Google analytics on
sagemath.org). From 2008 to 2011, year-on-year growth was about 50%,
which isn't great. However, from 2011 to now, year-on-year growth is
slightly less than 0%. It was maybe -10% from 2013 to 2014.
Incidentally, number of monthly active users of sagemath.org is about
68,652 right now, but the raw number isn't as import as the
year-to-year rate of change.

I set an overall mission statement for the Sage project at the outset,
which was is to be a viable alternative to the Ma's. Being a "viable
alternative" is something that holds or doesn't for specific people.
A useful measure of this mission then is whether or not people use
Sage. This is a different metric than trying to argue from "first
principles" by making a list of features of each system, comparing
benchmarks, etc., like Bill H. suggests above.


2. Guiding policies:

Statement of policy: focus on undergraduate students in STEM courses
(science, tech, engineering, math)

Justification:

In order for Sage to start growing again, identify groups of people
that are not using Sage. Then decide, for each of these groups, who
might find value in using Sage, especially if we are able to put work
into making it easier for them to benefit from Sage. This is
something to re-evaluate periodically.
In itself, this is very generic -- it's what any software project that
wishes to grow should do. The interesting part is the details.

Some big groups of potential future users of Sage, who use Sage very
little now, include:

- employees/engineers in various industries (from defense contractors,
to finance, to health care to "data science").

- researchers in area of mathematics where Sage is currently not popular

- undergraduate students in STEM courses (science, tech, engineering, math)

I think by far the most promising group is "undergraduate students in
STEM courses". In many cases they use no software at all or are
unhappy with what they do use. They are extremely cost sensitive.
Open source provides a unique advantage in education because it is
less expensive than closed source software, and having access to
source code is something that instructors consider valuable as part of
the learning experience. Also, state of the art performance, which
often requires enormous dedicated for-pay work, is frequently not a
requirement.

3. Actions:

- (a) Make access to Sage as easy as possible.

- (b) Encourage the creation of educational resources (books,
tutorials, etc.) that make using Sage for particular courses as easy
as possible.

- (c) Implement missing functionality in Sage that is needed in
support of undergraduate teaching.

Justification:

Why don't more undergraduates use Sage? For the most part, students
use what they are told to use by their instructors. So why don't
instructors chose to use Sage? (a) Sage is not trivial to install
(in fact it is incredibly hard to install), (b) There are limited
resources (books, tutorials, course materials, etc.) for making using
Sage really easy, (c) Sage is missing key functionality needed in
support undergraduate teaching.

Regarding (c), in 2008 Sage was utterly useless for most STEM courses.
However, over the years things changed for the better, due to the hard
work of Rob Beezer, Karl Dieter, Burcin Erocal, and many others.
Also, for quite a bit of STEM work, the numerical Python ecosystem
(and/or R) provides much of what is needed, and both have evolved
enormously in recent years. They are all usable from Sage, and
making such use *easier* should be an extremely high priority.
Related -- Bill Hart wrote "I recently sat down with some serious

developers and we discussed symbolics in Sage (which I know nothing
about). They argued that Sage is not a viable contender in that area,

and we discussed some of the possible reasons for that. " The reason
is that the symbolic functionality in Sage is motivated by making Sage
useful for undergraduate teaching; it has nothing to do with what
serious developers in symbolics would care about.

Regarding (b), an NSF (called "UTMOST") helped in this direction...
Also, Gregory Bard wrote "Sage for undergraduates", which is *exactly*
the sort of thing we should be very strongly encouraging. This is a
book that is published by the AMS and is also freely available. And
it squarely addresses exactly this audience. Similarly, the French
book that Paul Zimmerman edited is fantastic for France. Let's make
an order of magnitude similar resources along these lines! Let's
make vastly more tutorials and reference manuals that are "for
undergraduates".

Regarding (a), in my opinion the most viable option that fits with
current trends in software is a full web application that provides
access to Sage. SageMathCloud is what I've been doing in this
direction, and it's been growing since 2013 at over 100% year on year,
and much is in place so that it could scale up to more users. It
still has a huge way to go regarding user friendliness, and it is
still *losing money every month*. But it is a concrete action toward
which nontrivial effort has been invested, and it has the potential to
solve problem (a) for a large number of potential STEM users.
College students very often have extremely good bandwidth coupled with
cheap weak laptops, so a web application is the natural solution for
them.

Though much has been done to make Sage easier to install on individual
computers, it's exactly the sort of problem that money could help
solve, but for which we have little money. I'm optimistic that
OpenDreamKit will do something in this direction.

-- William


[1] I've attached plots of "active monthly users" for sagemath.org
and cloud.sagemath.com. Which is which should be clear.

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

[Bill Hart]

I don't disagree. But none of that is realistic.

I think $100k annual is more realistically the market rate. But then you get to be pushed around every day, have 9-5 working hours and deadlines. I've turned down offers to interview for such positions because I prefer the flexibility of academic work.

There's no chance of tenure at any academic institution if you are employed as a software engineer. So that's unrealistic.

And we are a mathematics dept. So hiring someone as a mathematics postdoc isn't going to get this job done.

I do accept the premise that higher compensation might feasibly attract applicants. It's not possible though. It looks to me like the position may go unfilled.

Bill.

[Dima Pasechnik]

On Wednesday, September 30, 2015 at 11:34:06 AM UTC-7, Bill Hart wrote:

I don't disagree. But none of that is realistic.

I think $100k annual is more realistically the market rate. But then you get to be pushed around every day, have 9-5 working hours and deadlines. I've turned down offers to interview for such positions because I prefer the flexibility of academic work.

There's no chance of tenure at any academic institution if you are employed as a software engineer. So that's unrealistic.

I was a research programmer for a year at Utrecht University (on CGAL); IMHO 

it didn't hurt my tenure prospects (not that I have a tenure now, but I'm still around in academia :-))

And it was with 9-5 (and more) working hours and deadlines, and I was pushed around quite a bit.

And we are a mathematics dept. So hiring someone as a mathematics postdoc isn't going to get this job done. 

No, why? Are you serious? Are you yourself a tenured professor now?

I do accept the premise that higher compensation might feasibly attract applicants. It's not possible though. It looks to me like the position may go unfilled.

you have to re-package it as a postdoc; after all it's probably publishable work, to get something working the way you want...

Or/and offer part-time and/or remore work. 

Dima

[Jeroen Demeyer]

On 2015-09-30 20:46, Dima Pasechnik wrote:
> after all it's probably publishable work

Unfortunately, I very much doubt that this is true. Developing
algorithms on paper is publishable, actually implementing them usually
isn't. And that's only if you're working on some algorithmic aspect.
Just making software work (porting, fixing bugs, ...) certainly isn't
publishable.

Jeroen.

[Bill Hart]

That's an interesting read William, and I think a pretty good aim and justification.

I realise that users of sagemath.org is a proxy for the number of users of Sage. It's no doubt better than the number of downloads, since many users might stick with an old version of Sage, be getting it from their distribution, or using it on cloud.sagemath.com. But it is also subject to at least some adjustment due to the availability of resources other than sagemath.org, such as the excellent resources you mentioned, the existence of stack overflow, courses using Sage, etc.

So growth may not have actually been negative. Users might simply be getting their information in more diverse locations.

Having said that, it does seem clear that growth has at least stagnated to some degree, though Sage remains about as popular as it has ever been.

The main questions that come into my mind after reading your post are:

* Out of curiosity, is there any indication that the seasonal fluctuation in usage of sagemath.org is largely or in part down to stem users?

* What are stem undergraduate students currently using predominantly? I'm guessing Mathematica, Maple and Matlab, but not Magma?

* What can Sage offer that disrupts what is being offered by the other contenders? Price seems to be one answer. But I seriously question whether 45 Euros a year (for a Mathematica student license) is a deal-breaker for most students (it's only the price of another textbook). And I'm sure they've done extensive research to set that value at the most they can ask for regionally to maximise usage and profit.

* How does the strategy feed the future development of Sage? Is the strategy to increase the number of competent Sage users in stem so that the number of competent developers, plus the number of people who need Sage when they enter the workforce is higher in the future?

Bill.

[Dima Pasechnik]

On Wednesday, September 30, 2015 at 11:51:35 AM UTC-7, Jeroen Demeyer wrote:

On 2015-09-30 20:46, Dima Pasechnik wrote:
> after all it's probably publishable work
Unfortunately, I very much doubt that this is true. Developing
algorithms on paper is publishable, actually implementing them usually
isn't.

there are many areas where benchmarks are a must. That is, one needs some

kind of implementation, preferably a fast one.

 

And that's only if you're working on some algorithmic aspect.
Just making software work (porting, fixing bugs, ...) certainly isn't
publishable.

this is true to a considerable extent, given that software is already written...

 

Jeroen.

[Bill Hart]

On Wednesday, 30 September 2015 20:46:29 UTC+2, Dima Pasechnik wrote:

On Wednesday, September 30, 2015 at 11:34:06 AM UTC-7, Bill Hart wrote:

I don't disagree. But none of that is realistic.

I think $100k annual is more realistically the market rate. But then you get to be pushed around every day, have 9-5 working hours and deadlines. I've turned down offers to interview for such positions because I prefer the flexibility of academic work.

There's no chance of tenure at any academic institution if you are employed as a software engineer. So that's unrealistic.

I was a research programmer for a year at Utrecht University (on CGAL); IMHO 

it didn't hurt my tenure prospects (not that I have a tenure now, but I'm still around in academia :-))

And it was with 9-5 (and more) working hours and deadlines, and I was pushed around quite a bit.

Sorry to hear that. We don't push people around at my institution.

 

And we are a mathematics dept. So hiring someone as a mathematics postdoc isn't going to get this job done. 

No, why? Are you serious? Are you yourself a tenured professor now?

Because a mathematics postdoc is expected to publish mathematics, not write software. This is part of the reason the ODK project has focused so much on software engineers. We need them, but mathematics postdocs have entirely different skill sets and aspirations.

 

I do accept the premise that higher compensation might feasibly attract applicants. It's not possible though. It looks to me like the position may go unfilled.

you have to re-package it as a postdoc; after all it's probably publishable work, to get something working the way you want...

Umm what!?

 

Or/and offer part-time and/or remore work. 

We can't do that locally. I've asked.

Bill.

[Volker Braun]

On Wednesday, September 30, 2015 at 8:34:06 PM UTC+2, Bill Hart wrote:

I think $100k annual is more realistically the market rate.

http://www.glassdoor.com/Salary/Google-Salaries-E9079.htm

And I don't think that the average software engineer can write a superoptimizer that beats gmp/mpir in a year. Excluding any google engineers on this list, of course ;-)

Its true that salaries in Germany are a bit lower, but then limiting yourself to the ones that don't want to move isn't exactly helpful if you want them to be in Kaiserslautern.

I'm not saying that its your fault, its more of a systemic problem in Mathematics. Either you provide a career path for scientists writing software, or you pay a lot more for the private sector to do it for you.

[Bill Hart]

On Wednesday, 30 September 2015 21:19:08 UTC+2, Volker Braun wrote:

On Wednesday, September 30, 2015 at 8:34:06 PM UTC+2, Bill Hart wrote:

I think $100k annual is more realistically the market rate.

http://www.glassdoor.com/Salary/Google-Salaries-E9079.htm

64,000 Euros, 103,000 Euros, 88,000 Euros. Those are the salaries Google is offering. Looks like I was pretty close.

 

And I don't think that the average software engineer can write a superoptimizer that beats gmp/mpir in a year.

I know someone who did it in 3 months. I'm quite certain it can be done in a year, especially if I teach them how to do it and especially if that is their full time job. :-)

 

Excluding any google engineers on this list, of course ;-)

Its true that salaries in Germany are a bit lower, but then limiting yourself to the ones that don't want to move isn't exactly helpful if you want them to be in Kaiserslautern.

I'm not saying that its your fault, its more of a systemic problem in Mathematics. Either you provide a career path for scientists writing software, or you pay a lot more for the private sector to do it for you.

Yeah I can't disagree with you on that last point.

Bill. 

[Volker Braun]

On Wednesday, September 30, 2015 at 9:28:14 PM UTC+2, Bill Hart wrote:

http://www.glassdoor.com/Salary/Google-Salaries-E9079.htm

64,000 Euros, 103,000 Euros, 88,000 Euros. Those are the salaries Google is offering. 

You followed the redirect to the German page. As I said, salaries in Germany are lower but then you are limiting yourself to those that don't want to move. Actual average is listed as $162,132

[Bill Hart]

Here is Europe, the student pricing to buy a copy of Maple outright is even cheaper than Mathematica. And they also offer options for institutions to offer Maple free to their students.

Moreover, Maple offer a way for students to use Maple worksheets provided by their professors for free.

Matlab is so cheap for students it may as well be free. (Their strategy is surely to get students using it for next to nothing so that they will demand it when they enter the workforce.)

Bill.

[Bill Hart]

On Wednesday, 30 September 2015 21:36:16 UTC+2, Volker Braun wrote:

On Wednesday, September 30, 2015 at 9:28:14 PM UTC+2, Bill Hart wrote:

http://www.glassdoor.com/Salary/Google-Salaries-E9079.htm

64,000 Euros, 103,000 Euros, 88,000 Euros. Those are the salaries Google is offering. 

You followed the redirect to the German page.

I imagine it redirected automatically since I am in Germany.

 

As I said, salaries in Germany are lower but then you are limiting yourself to those that don't want to move. Actual average is listed as $162,132

It's extremely unlikely we'll get someone wanting to move to the EU from overseas for a position only guaranteed for one year. We are at least being realistic about it.

Bill. 

[Jeroen Demeyer]

On 2015-09-30 21:11, Dima Pasechnik wrote:
> On Wednesday, September 30, 2015 at 11:51:35 AM UTC-7, Jeroen Demeyer wrote:
>
> On 2015-09-30 20:46, Dima Pasechnik wrote:
> > after all it's probably publishable work
> Unfortunately, I very much doubt that this is true. Developing
> algorithms on paper is publishable, actually implementing them usually
> isn't.
>
> there are many areas where benchmarks are a must. That is, one needs some
> kind of implementation, preferably a fast one.

"some kind of implementation" is very different from an implementation
which is user friendly, portable, easy to compile & install, bug-free,...

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 30 September 2015 at 18:05, Jori Mäntysalo <Jori.Ma...@uta.fi> wrote:

It is usually good to build on strong points. So, what are best areas in Sage? Where it now is The Software(tm) to use?

And how could we expand those to some near area?

I'm not convinced that is true - in fact, I would go as far as to say it is false.

Will adding more number theory stuff into Sage increase the user base? I suspect the fact that Sage is already strong in that area, means it will not.

Would improving the symbolic maths help? I suspect it would, as rightly or wrongly, (and I don't claim to be a good judge of that), I think Mathematica is seen by many as the best overall package for symbolic maths.

I used vector network analyzers (VNAs) a lot in my job. They measure the impedance of devices, and normally save data as S-parameters (Scattering parameters), which basically have frequency in one column, and the magnitude and angle of a reflected or transmitted wave in other columns. The file format is pretty simple - here's a bit of an example file, which I collected from one of my instruments using the GPIB bus.

! This is a touchstone format file.
! It should be saved with .s2p extension
! HEWLETT PACKARD,8720D,0,7.74 
! Date = 31 Aug 2013
! Time = 18:37:05
! Start frequency = 0.050000000 GHz
! Stop frequency  = 6.000000000 GHz
! IF bandwidth = 100 Hz
! Averaging = OFF
! Averaging factor = 2
! Port extensions = OFF
! Port extension 1 = 0.000000 ps
! Port extension 2 = 0.000000 ps
! Points  = 1601
! Calibration kit = User-defined
! Calibration = Full 2-port
! freq magS11 angS11 magS21 angS21 magS12 angS12 magS22 angS22
! Magnitudes are in log form
! Angles are in degrees
# MHz S DB R 50
50 -61.8945 -16.5234 -29.7334 -2.63403 -29.7471 -2.63953 -60.0195 39.3203
53.7188 -63.7793 -20.7686 -29.7539 -2.89636 -29.7461 -3.00757 -59.5742 37.5234
57.4375 -62.4453 -4.62256 -29.7158 -3.19153 -29.7617 -3.09814 -59.7539 34.6055
61.1562 -60.2402 -1.19067 -29.7373 -3.16272 -29.7656 -3.47717 -60.5273 37.7832

Now since it contains data as a function of frequency, one can get it as a function of time by doing an Inverse Fourier Transform (IFT). Now Keysight, who produce a lot of VNAs charge you around $3000 to do enable the software in their instruments to do the IFT and so display data as a function of time. Now lots of people don't have that option, so it is not uncommon to see people wanting to do the IFT outside their instrument.

I normally direct them to bit of free software, that is designed for a low cost (<$1000) vector network analyzer, but the author, who is an academic, has written it in such a way that it can read the above file format and display the time-domain data.

I'm sure it would not be rocket science in Sage to do this, but you would have to write code to parse the file properly, do the IFT with windowing. It is much easier to just use a bit of software that has the functionality built in.

Dave

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 30 September 2015 at 20:44, Bill Hart <goodwi...@googlemail.com> wrote:

It's extremely unlikely we'll get someone wanting to move to the EU from overseas for a position only guaranteed for one year. We are at least being realistic about it.

Bill. 

Certainly many engineers would do it.  A contract post for a year would suite many, but they would expect to earn a reasonable amount of money out of that.

Dave

[Bill Hart]

I don't see why would you make people use Sage for that, just so you can get the Sage usage figures up, instead of writing that as a simple Python package.

Just taking other software and making it part of Sage so that "Sage can do that too" isn't an argument for the adoption of Sage.

I think I see that you would like a community of electronics engineers in the Sage community. But I'm not personally convinced that is the way to do it.

I think if it were me I'd set up a sage-engineering subproject and start writing code that uses Sage, starting with the things that the greatest number of engineers need in a very broad sense. Then figure out what Sage can offer them that other projects don't, and then build it. But I'm not an electronics engineer, so I really have little insight into that.

Bill.

[Bill Hart]

The wages at Universities in Germany are fixed based on qualifications and number of years experience. We have absolutely no say in the salaries whatsoever.

In the end, if we find someone for the position, it will be because they specifically want to come and work with our research group, and because they are very interested in superoptimisers/assembly language. The sort of candidate we are looking for isn't likely to be someone currently looking for work for a high salary in industrial engineering. The MPIR position is open to someone with an undergraduate degree, preferably in maths, computer science or electrical engineering (with computer specialisation).

So far, people have offered to do it remotely, or without a degree. But it is essentially impossible for us to accommodate that.

There is definitely the possibility that someone currently working in a Laboratoire in France on say GPU optimisation, someone who has just finished a degree in electrical engineering, working on chip design or embedded hardware, etc. or someone who has tried the industrial life but not found it suitable for them, will apply.

Bill. 

[Michael Orlitzky]

On 09/30/2015 02:25 PM, William Stein wrote:
> However, from 2011 to now, year-on-year growth is
> slightly less than 0%. It was maybe -10% from 2013 to 2014.

There are probably a lot of people like me who haven't been back to
sagemath.org since the switch to git. I just `git pull` every once in a
while.

[Travis Scrimshaw]

Sometimes I don't even tell people to go there, but instead just tell them the necessary git commands.

Best,
Travis

[William Stein]

It's a fair argument that "monthly active users of sagemath.org" is
not exactly the same thing as "users of sage", or something else.
However, in my experience, when I look at pretty much any measure of
the growth of a product in a marketplace -- it doesn't really matter
which -- they tend to be highly correlated. I'm not making any
claims at all about exact numbers of "users of sage". I'm talking
only about year-over-year growth of a well-defined metric. The
metric "monthly active visitors to the sagemath.org" website has the
advantage that it is well defined and we have data about it going back
to before 2008, so it is useful. I'm concerned because here was
basically no year-over-year growth of that
metric since 2011.

This lack of growth should be a serious cause of concern for us Sage
developers. It is concrete numerical data that complements Bill
Hart's concerns expressed
elsewhere: "There's nothing I can see technologically in Sage that

even has the potential to disrupt Magma's dominance. Not even in
theory."

Sage is very good and useful in many ways to us. However, just as
Bill points out, there are very, very real difficulties, and thinking
about them strategically -- which is what Bill suggested we do above
-- is a really good suggestion. He made a list of questions, and I've
also shared my current strategy above (basically: undergraduate STEM
education!). I hope people will support my suggestion, and be as
supportive as possible of efforts to improve Sage to make it better
for undergraduate teaching -- let's try to listen to and appreciate
what people like Karl Dieter Crisman, Rob Beezer, Gregory Bard, and
Paul Zimmerman, and these other people suggest carefully, since they
have tons of classroom experience and great ideas. Let's make new
documentation like Gregory Bard's "Sage for undergraduates" book.

In the long run if Sage is used 1000 times as much in undergraduate
classrooms, the Sage project will be much better off, though this
might sometimes mean making painful design decisions that are less
supportive of
research.

-- William

>
> --
> You received this message because you are subscribed to the Google Groups "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
> To post to this group, send email to sage-...@googlegroups.com.
> Visit this group at http://groups.google.com/group/sage-devel.
> For more options, visit https://groups.google.com/d/optout.



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

[Jonathan]

As a physical scientist (chemist) who does research and teaches at the undergraduate level I agree with William that STEM is probably a place where Sagemath could be very useful.

However, I'm not sure the reasons are what most people think.  Before I try to explain myself, let me explain how presently I use Sagemath in my teaching:
1) I primarily use it as a tool to help students more rapidly do messy symbolic manipulations (most involving integration and differentiation of mathematical models of physical systems).  In general the students then need to take the symbolic result and plot it over a physically interesting range or compute a numerical result for specific conditions.
2) I also use tutorials that I built using python within Sagemath to help students learn to convert physical situations into their mathematical representations.  A particular example that I use regularly is writing the system of differential equations that represents a complex multi-step chemical reaction.
3) There are other examples, but one that I do not use Sagemath for is analysis of large data sets (we use IGOR, by Wavemetrics, or LoggerPro, by Vernier) because inputting large columns of data using copy and paste (what students will use effectively) does not work well.  These other programs also easily make publication quality scientific plots.

So why do I think Sagemath is good for STEM?
1) MOST IMPORTANTLY it is open source.  This is important because science depends on anybody in the future being able to reproduce your work and for current workers to check that what they are doing is done the way they think it is.  Thus commercial (black box) software is difficult to use for good science.  If you do a particular analysis of data in a commercial piece of software and that software becomes unavailable or will not run on the machines of another worker then your work is not reproducible.  Although it may be difficult, open-source software is more future proof and platform agnostic than commercial compiled code.  I make a point of this to my students.  If they get a weird result from an open source piece of software they can examine exactly what the code is doing; thus either finding a bug in the code or their procedures or data.  With commercial software you are just stuck with the result.  That said, I do use commercial software, but carefully limit myself to software that documents exactly how every algorithm is implemented (IGOR references code from _Numerical_Recipes_). In short open source is good science.

2) Science uses a huge array of different mathematical manipulations and representations, thus a project like Sagemath that includes such a wide array of math tools is very useful.

3) Python is used by a lot of the scientific community for manipulation of data files (primarily text representations).  Thus Sagemath is a way of introducing students to this.  I initially get them hooked by providing a tool that removes some of the drudgery and errors associated with student algebra and calculus manipulations.

4) There are lots of places where Sagemath falls short for STEM use (copy and paste of large data sets, high quality zoomable 2-D graphics, simple tools for computations with significant figures, a units package that handles units the way a physical scientist or engineer does and other problems).  However, I have found it easier to teach to my students than Maple (also available on my campus) and cheaper for them long term.

I could list more reasons I think Sagemath could be very important to STEM, but #1 is the most important.

Jonathan

[Jori Mäntysalo]

On Wed, 30 Sep 2015, Dr. David Kirkby (Kirkby Microwave Ltd) wrote:

>> It is usually good to build on strong points. So, what are best areas
>> in Sage? Where it now is The Software(tm) to use?

>> And how could we expand those to some near area?

> I'm not convinced that is true - in fact, I would go as far as to say it
> is false.
>
> Will adding more number theory stuff into Sage increase the user base? I
> suspect the fact that Sage is already strong in that area, means it will
> not.

OK, number theory is a strong point. I do not think about doing more
number theory. Is there some *near* field of mathematics, where Sage would
be quite good but not excellent?

(If growing wheat is a strong point, next try oak - no carrots.)

--
Jori Mäntysalo

[Dima Pasechnik]

come, come. Implementing something mathematically non-trivial on a GPU for the 1st time is still

publishable research, although not in mathematics, but somewhere in between mathematics and CS.

See e.g. Communications of ACM to see highlights of this sort of research.

And most people who do good implementations appreciate the maintainability of code they write.

  

[Dima Pasechnik]

On Wednesday, 30 September 2015 12:12:51 UTC-7, Bill Hart wrote:

On Wednesday, 30 September 2015 20:46:29 UTC+2, Dima Pasechnik wrote:

On Wednesday, September 30, 2015 at 11:34:06 AM UTC-7, Bill Hart wrote:

I don't disagree. But none of that is realistic.

I think $100k annual is more realistically the market rate. But then you get to be pushed around every day, have 9-5 working hours and deadlines. I've turned down offers to interview for such positions because I prefer the flexibility of academic work.

There's no chance of tenure at any academic institution if you are employed as a software engineer. So that's unrealistic.

I was a research programmer for a year at Utrecht University (on CGAL); IMHO 

it didn't hurt my tenure prospects (not that I have a tenure now, but I'm still around in academia :-))

And it was with 9-5 (and more) working hours and deadlines, and I was pushed around quite a bit.

Sorry to hear that. We don't push people around at my institution.

 

And we are a mathematics dept. So hiring someone as a mathematics postdoc isn't going to get this job done. 

No, why? Are you serious? Are you yourself a tenured professor now?

Because a mathematics postdoc is expected to publish mathematics, not write software. This is part of the reason the ODK project has focused so much on software engineers. We need them, but mathematics postdocs have entirely different skill sets and aspirations.

 

I do accept the premise that higher compensation might feasibly attract applicants. It's not possible though. It looks to me like the position may go unfilled.

you have to re-package it as a postdoc; after all it's probably publishable work, to get something working the way you want...

Umm what!?

call these jobs WIMI (that's the right German abbreviation for a research fellow, IIRC), and not engineer.

And let these people continue doing publishable research on their own, if they wish, for 20-25% of their time.

And ask them to publish on stuff they'd get to work on in your project, not merely code.

This way you might have a fighting chance to get  CS/maths people who are potentially interested 

in academic career in CS/maths.

IMHO, it's hopeless to attract people who are just interested in a career in industry with what you offer.

[Jeroen Demeyer]

On 2015-10-01 04:13, Jonathan wrote:
> high quality zoomable 2-D graphics
It's interesting that you think that Sage cannot do this. Because I
think that Matplotlib's plots (which is what Sage uses) are "high
quality zoomable 2-D graphics". So I wonder what you're missing.

[Volker Braun]

On Thursday, October 1, 2015 at 8:13:05 AM UTC+2, Dima Pasechnik wrote:

call these jobs WIMI (that's the right German abbreviation for a research fellow, IIRC), and not engineer.

"Wissenschaftlicher Mitarbeiter"

Still, one year is not enough to do the work, publish, and use that for the next application round. So its carer suicide. Ok we've already established that this might only appeal to people without career ambitions.

If you think the work can be done in 3 months then hire a contractor. 90 * 600 is probably still below your budget.

[Juan Luis Cano]

Hi all,

Francesco's words resonated a lot with me, and I felt like telling my own story in case it is of any interest.

I started my MSc in Aerospace Engineering in 2009 and got into open source math software trough Sage around 2010 (probably googling "free alternative to Mathematica"). Actually I did some experiments with the notebook and explored the idea of implementing it using websockets[1]. At some point though, I was discouraged by the difficulties of installing Sage and the syntax differences from pure Python, so I started exploring NumPy, SciPy, matplotlib and SymPy outside of Sage. These libraries essentially covered everything I needed, were infinitely easier to install and the community was bigger - hence I had lots of tutorials and StackOverflow answers.

In the meanwhile, I have been organizing the Python Conference in Spain (PyConES) since 2013 and every year there is an "Introduction to Sage" talk which essentially presents extremely basic functionality already available in IPython/Jupyter with plain NumPy and matplotlib. And I find myself thinking "Why do they advertise this instead of the true selling point of Sage: pure math?" I've been giving talks and courses in Europe and Latin America and find myself happy without Sage. I feel sorry!

The situation got better over the years (the IPython notebook was born, then Anaconda came to scene so I never ever had to recompile NumPy again) and it's been ages since I last used (and recommended) Sage for any real work. Perhaps Sage symbolics are better than SymPy - but what's the cost? I do not know a single classmate who would find in Sage a key functionality missing in the "SciPy stack". Let me stress again that we are engineers - most of what we do is solving ODEs, plotting and reading from files.

I have been reading very closely all William posts about the topic and am honestly worried about Sage's future, because it's a project I still love. Its developers have poured a tremendously big amount of top-quality, unpaid work. But I agree with others that the project diversified too much and maybe should start *dropping* things instead of *adding* things. What about "Being a viable alternative to Magma" - and leave the rest of the Ma's out? What about removing something of those 800 Mb? Splitting core functionality and making it available through pip and conda on PyPI? Stating it as a federation of projects? Focus on its strengths and its present community (that is: pure mathematics) rather than pleasing everybody?

Sage's current mission statement is epic and desirable, but maybe it is too broad.

My two cents.

Juan Luis Cano

[1] https://groups.google.com/d/msg/sage-notebook/JbJSULEX3hA/6N73HltFFEEJ

On Tuesday, September 29, 2015 at 4:44:07 PM UTC+2, bluescarni wrote:

On 28 September 2015 at 19:37, William Stein <wst...@gmail.com> wrote:

1. Magma is also an Ma.   Magma's incredibly good at pure mathematics.
You seem to be leaving out Magma above.

I admit I know basically nothing about Magma (I did not know it even existed before joining this list :).

 

2. You say "... better served in the Python space by the
Numpy/SciPy/SymPy/Matplotlib stack as an alternative to the Ma's
rather than SAGE."   Sage includes "Numpy/SciPy/SymPy/Matplotlib", so
we don't have to worry about that chunk of people with respect to our
mission statement.

I can install the NSPM stack on any modern platform (Windows included) with pip (or my distro's package manger, if I am on linux) in probably less than 5 minutes with a reasonable internet connection. I don't have to worry about sizeable downloads, virtual machines, containers, emulation layers a-la cygwin, installing a separate compiler toolchain/python version/set of libraries, or anything of the sort.

It is true that you can use the NSPM stack from SAGE, but what are the key advantages of doing so? It is a honest question, maybe there's something I am overlooking.

3. There is a lot more to mathematics than just what Magma does and
*also* much more to it than just what Numpy/SciPy/SymPy/Matplotlib do.
  There's a huge amount of interesting things that could be
systematically computed with in mathematics that no existing package
does yet.

Yes, nobody has the monopoly on what "mathematics on a computer" means :)

My comment was merely a marketing/strategic one: I think there exists a disconnect between the mission statement and what SAGE actually is (and maybe what the SAGE community wants it to be).

Cheers,

  Francesco.

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 30 Sep 2015 21:51, "Bill Hart" <goodwi...@googlemail.com> wrote:
>
> I don't see why would you make people use Sage for that, just so you can get the Sage usage figures up, instead of writing that as a simple Python package.
>

But if Sage could do a lot of the things related to this, which RF engineers can do with the MATLAB toolbox, then Sage would start  becoming a viable alternative to MATLAB for engineers working in this field. At the moment, Sage is nowhere near a viable alternative, and I don't think it will in my lifetime (I'm 51), if ever.

If you look at Mathematica for example, it has evolved a lot, to add functionality in areas far removed from version 1.0. Apparently the biggest user base is actually the financial sector - or at least was a few years ago. (This possibly hints at another source of funding for Sage - the financial sector. )

I think to be honest, there's a good argument for just re-writing the Sage "Mission Statement", since realistically the mission has zero chance of ever being reached, or even approached fairly closely. I don't think you could come up with any very objective metrics, but I believe the gap between Mathematica and Sage is widening, as it the gap between Sage and MATLAB.

If I am honest, I think the *only* way Sage would ever be a viable alternative to Mathematica for a very large number of users would be if Wolfram Research stopped development of the program, either because they went bust, or decided it was not commercially viable, so stopped development of it voluntarily. The same argument would apply to MATLAB with Mathworks.

Dave

[Bill Hart]

I suspect you are right, that for large classes of user, Sage will never catch these multimillion dollar companies. They aren't static targets and they sure as hell aren't going to allow Sage to muscle its way in.

Perhaps William is right in trying to focus on making Sage a viable alternative for one particular kind of user: stem undergrads. That is probably an achievable and worthwhile goal.

(I'm sure he'd also like it to be a viable alternative for academic mathematicians too, since he is one himself. But quite obviously he's saying that in the design of Sage certain decisions will need to be made if his goal is made a focus.)

Bill.

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 30 Sep 2015 21:51, "Bill Hart" <goodwi...@googlemail.com> wrote:
>

> I don't see why would you make people use Sage for that, just so you can get the Sage usage figures up, instead of writing that as a simple Python package.
>

But if Sage could do a lot of the things related to this, which RF engineers can do with the MATLAB toolbox, then Sage would start  becoming a viable alternative to MATLAB for engineers working in this field. At the moment, Sage is nowhere near a viable alternative, and I don't think it will in my lifetime (I'm 51), if ever.

If you look at Mathematica for example, it has evolved a lot, to add functionality in areas far removed from version 1.0. Apparently the biggest user base is actually the financial sector - or at least was a few years ago. (This possibly hints at another source of funding for Sage - the financial sector. )

I think to be honest, there's a good argument for just re-writing the Sage "Mission Statement", since realistically the mission has zero chance of ever being reached, or even approached fairly closely. I don't think you could come up with any very objective metrics, but I believe the gap between Mathematica and Sage is widening, as it the gap between Sage and MATLAB.

If I am honest, I think the *only* way Sage would ever be a viable alternative to Mathematica for a lot of users would be if Wolfram Research stopped development of the program,. either because they went bust, or decided it was not commercially viable, so stopped development of it voluntarily. The same argument would apply to MATLAB with Mathworks.

Dave

[Bill Page]

On 1 October 2015 at 05:47, Juan Luis Cano <juan...@gmail.com> wrote:
> ...

> The situation got better over the years (the IPython notebook was born, then
> Anaconda came to scene so I never ever had to recompile NumPy again) and
> it's been ages since I last used (and recommended) Sage for any real work.
> Perhaps Sage symbolics are better than SymPy - but what's the cost? I do not
> know a single classmate who would find in Sage a key functionality missing
> in the "SciPy stack". Let me stress again that we are engineers - most of
> what we do is solving ODEs, plotting and reading from files.
>

+1

There is one original feature unique to Sage however that originally
attracted me to Sage in the very early years and which continues to
motivate my interest. This originally was summed up in a seemingly
trite slogan like: "Building the car, not re-inventing the wheel" but
which no longer seems very evident. That is the Python-based "glue"
that makes it possible for very different "external" packages such as
Maxima, Gap, Singular, Octave and many others (including especially in
my case FriCAS) to appear as objects within a universal interface in
Sage. I have been disappointed about the apparent shift in Sage away
from this goal.

On the other hand I am very encouraged by the continued development of
SMC which makes exactly the kind of transition that Juan describes
very easy while retaining the "glue" implemented in Sage.

> I have been reading very closely all William posts about the topic and am
> honestly worried about Sage's future, because it's a project I still love.
> Its developers have poured a tremendously big amount of top-quality, unpaid
> work. But I agree with others that the project diversified too much and
> maybe should start *dropping* things instead of *adding* things. What about
> "Being a viable alternative to Magma" - and leave the rest of the Ma's out?

As a core developmental strategy for Sage this seems like a reasonable
goal to me (although I do not anticipate any immediate need for this
kind of functionality in my own work). Maybe this is a logical
extension of the combinat work on the "category framework" in Sage.
Unfortunately while I am very much in favor of the category/domain
approach of Axiom and related systems, I find the Sage implementation
of this idea almost entirely indigestible. Perhaps this is not the
case for a sufficiently large number of potential Sage developers.

I am sorry, but I do not think an emphasis on STEM-oriented features
in Sage is likely to have a big impact on it's use or development. It
could however it could potentially contribute a lot to the wider
adoption of SMC and I agree that that would be a good thing.
Unfortunately there are already a number of both open and proprietary
competitors in this highly commercialized and commercializable field.

> What about removing something of those 800 Mb? Splitting core functionality
> and making it available through pip and conda on PyPI? Stating it as a
> federation of projects? Focus on its strengths and its present community
> (that is: pure mathematics) rather than pleasing everybody?
>

+1

To me "federation" is exactly what Sage was *originally* about.

[Bill Hart]

On Thursday, 1 October 2015 16:35:20 UTC+2, Bill Page wrote:

<SNIP>

Unfortunately while I am very much in favor of the category/domain
approach of Axiom and related systems, I find the Sage implementation
of this idea almost entirely indigestible.  Perhaps this is not the
case for a sufficiently large number of potential Sage developers.

 Bill, I would be very interested if you could elaborate on this point in more detail (assume I don't know anything about Aldor/Spad because it is so long ago that I read the manual for Aldor that I really have forgotten how this works over there).

I'm currently implementing a system in Julia which really follows the design of Sage's parent/element setup quite closely (we tried another approach, but it failed). Since the project I'm working on is yet very young, I'd like to understand what is unpalatable about the Sage approach, since I'm essentially using it.

Bill. 

[Dima Pasechnik]

On Thursday, 1 October 2015 01:41:03 UTC-7, Volker Braun wrote:

On Thursday, October 1, 2015 at 8:13:05 AM UTC+2, Dima Pasechnik wrote:

call these jobs WIMI (that's the right German abbreviation for a research fellow, IIRC), and not engineer.

"Wissenschaftlicher Mitarbeiter"

Still, one year is not enough to do the work, publish, and use that for the next application round. So its carer suicide. Ok we've already established that this might only appeal to people without career ambitions.

Bill, 

if you think that ODK money is going to waste this way, perhaps you might consider giving it to another ODK party, to hire someone to do work for  you remotely.

Incidentally, my HR appears to be telling me that they made a clerical error, and as of today the only funding I have is the ODK funding, and it's 50% time funding (that's how much

I asked from ODK). They sent me a letter last year, that appeared to be a contact extension, and they say it was a typo in the date. I am still trying to see what this means...

Not that I can do a super-duper GPU programming for you (at least not immediately), but at least I can write and debug C code :-)

Dima

[Peter Howe]

I'd like to wholeheartedly agree with Jonathan.

First, free software: I work in chemistry for a large multinational company, and Sage has been really useful for me.  I've used both Mathematica and Matlab there, but have dropped them - it is really hard to persuade managers to cough up the licence fees for on-going support unless you can point to a development project that will fail without them.  The fees are eye-watering - typically 20% of purchase cost.  Even if you do buy a copy and find the money for support, you can then only install it on one PC at work.  With Sage, I can access in multiple places, including my home PC for working at home.

Second, the sophistication of maths in Sage.  I dropped maths at 18, so I'm not doing anything sophisticated but the way I can do algebra, data manipulation and plotting in the same environment is great.  I'm dealing with simple equations and small data sets (tens of points/<5 variables).  But I don't need to chop and change between programs like with Matlab/Mathematica.  The lack of sophistication needed by some of the folks here doesn't apply to me - or, I think, many chemists/biochemists.

Third, web-based.  Companies have large, paranoid IT departments so just downloading and installing stuff on a desktop can be impossible.  Getting licence terms agreed for commercial software can takes weeks, and then you have to get the install authorised and done by IT support (the only ones with admin rights).  Web is the way to go because it avoids those complications and makes it easy to share work and resource with colleagues - and you can probably get it squeezed onto a fast server which was created for some other application.

Finally, Python.  Learn Python and you are learning programming, unlike Mathematica or Matlab.  It is a decent language with consistent syntax and it makes automating tasks so easy (last week, I added leave-n-out cross-validation to a fitting program in 20 minutes; I mention it because I'd expected it to take half a day).  When you've learned Python in Sage, you can use it elsewhere!

The downsides?  Well, the Sage install is huge so that is off-putting, loading modules like matplotlib and R isn't intuitive and getting data in is a pain as others have commented.  A simple CSV upload utility would be nice...

Those are just my perspectives.  I agree completely with William that SMC is the way to go, because of the ease of access.  I think students may be an area to aim at because of a) Python b) Python c)  Python d) not having to install anything e) being able to do maths (differentiation etc) easily to understand the physical basis of many phenomena and then compare it with your results.  Symbolic maths was a revelation to me when I discovered it as a post-doc (!), and I think it still would be to many undergrads in chemistry/biochem/biology.  I'd love to have the time to put together some in interactive worksheets to go with textbooks, as some of the course tutors here have.

As also noted, the commercial Ma's do offer very very cheap subscriptions to students so I think that's less of a consideration.

Peter.

 

[Bill Hart]

On Thursday, 1 October 2015 19:40:30 UTC+2, Dima Pasechnik wrote:

On Thursday, 1 October 2015 01:41:03 UTC-7, Volker Braun wrote:

On Thursday, October 1, 2015 at 8:13:05 AM UTC+2, Dima Pasechnik wrote:

call these jobs WIMI (that's the right German abbreviation for a research fellow, IIRC), and not engineer.

"Wissenschaftlicher Mitarbeiter"

Still, one year is not enough to do the work, publish, and use that for the next application round. So its carer suicide. Ok we've already established that this might only appeal to people without career ambitions.

Bill, 

if you think that ODK money is going to waste this way, perhaps you might consider giving it to another ODK party, to hire someone to do work for  you remotely.

That's definitely something I hadn't considered. Of course deciding whether that is possible is out of my hands. Professors and HR depts only have a say about such things in Germany (not a problem for me, it saves me doing a lot of paperwork, as I'm not a professor). I will definitely run this past someone at KL in case we need to go down that route.

As an absolute last resort, it seems to me that giving the money to another ODK party rather than giving it back entirely is certainly more appealing. I'm still hopeful it won't come to that.

 

Incidentally, my HR appears to be telling me that they made a clerical error, and as of today the only funding I have is the ODK funding, and it's 50% time funding (that's how much

I asked from ODK). They sent me a letter last year, that appeared to be a contact extension, and they say it was a typo in the date. I am still trying to see what this means...

Ouch, that sounds pretty awful indeed. Do keep me updated on how that turns out.

 

Not that I can do a super-duper GPU programming for you (at least not immediately), but at least I can write and debug C code :-)

The main problem I can see would be if we were responsible for delivering the deliverables and it became difficult to communicate with someone working remotely. It's quite likely I will spend a considerable quantity of time explaining what I know and going through planning stages for that project with the person who ends up doing it. That's clearly easier to do in front of  whiteboard in my office than via email.

Anyway, as I say, I'm not able to make those sorts of decisions either way.

Bill.

[Jonathan]

When I say high-quality zoomable, I mean publication quality (see the kind of plots that appear in Science, American Scientist, Nature or any American Chemical Society or American Institute of Physics journals) and interactively zoomable by click and drag. To be useable for most people they also need GUI adjustment of colors, symbols, etc.  I started to work on some of these features for the 3-D graphics, but simply do not have the time necessary.  Anyway, I like Sage a lot, but we are very short on resources to bring things to the necessary level in the user interface.

The backend is good some places and not others, but is potentially very strong and robust because of the use of other packages.  As I track this discussion, I think that I like the idea of focusing on how to make all these things easily added and removed packages as necessary.  Maybe ipython/jupyter with good pypi packaging (I'm still not clear on how to package non-python stuff this way) is the way to go.  That would change Sagemath a lot, but might be a better use of resources.

Jonathan

Jonathan

[kcrisman]

When I say high-quality zoomable, I mean publication quality (see the kind of plots that appear in Science, American Scientist, Nature or any American Chemical Society or American Institute of Physics journals) and interactively zoomable by click and drag. To be useable for most people they also

Right, this is in principle possible but hasn't been implemented in the notebook yet.  William, I thought maybe you had made some pretty significant progress toward this in SMC?

[Jeroen Demeyer]

On 2015-10-02 04:01, Jonathan wrote:
> interactively zoomable by click and drag. To be useable for most people
> they also need GUI adjustment of colors, symbols, etc.

I saw some demo of a IPython/Jupyter widget which does exactly this: the
plot is essentially a Javascript applet which allows zooming client-side
(even without having Jupyter running).

[rjf]

I think that if you look back at the early rave reviews of Mathematica in such

notable scientific journals as the New York Times, you will see that the

reporters were impressed by color graphics of 3-d plots and endorsement

of such scientific notables as Steve Jobs, Also the eccentric aspects of

the humble and modest Stephen Wolfram.  Also, giving it away on

NeXt machines made an impression.   They did not notice that there

were (at the time) better graphics programs, better computers, and

even more eccentric people.  They seem to have lots of them coming

from British schools.

Most software developers seeking funding need a "killer app".  I don't

know that Mathematica has one -- but maybe it is STEM education,

since that's the major way of selling lots of systems.  There were

forays into financial software, engineering, visualization, web hosting,

information storage "curated"  (Alpha).  All of these were however

premised on the sale of the system to users, or selling of online

services  (or maybe ads?)

If you insist on free and open source, it seems to me that you

get nothing from your markets.  So a killer app does nothing for you.

(Is this the killer app du jour:  rapid dating ?)

Who was it who said (approximately) that "I spend my days writing

free software so that I can earn money in the evening delivering pizza"?

I suggested that the market for my (free, open source) TILU online integration

service could be the selling of advertising, presumably to freshman

calculus students.  ( soft drinks?  acne medication? )   but never

pursued it.  Maybe online ads for Sage???

Otherwise,I think you are left with finding and impressing one of those

rich people who thinks math programs are kool.

(You will have to judge for yourself how much of the above note

is sarcasm.)

Good luck.

[William Stein]

On Sat, Oct 3, 2015 at 8:38 AM, rjf <fat...@gmail.com> wrote:
> I think that if you look back at the early rave reviews of Mathematica in
> such
> notable scientific journals as the New York Times, you will see that the
> reporters were impressed by color graphics of 3-d plots and endorsement
> of such scientific notables as Steve Jobs, Also the eccentric aspects of
> the humble and modest Stephen Wolfram. Also, giving it away on
> NeXt machines made an impression. They did not notice that there
> were (at the time) better graphics programs, better computers, and
> even more eccentric people. They seem to have lots of them coming
> from British schools.
>

Hi -- thanks for sharing the above; I've never seen those articles.

> Most software developers seeking funding need a "killer app". I don't
> know that Mathematica has one -- but maybe it is STEM education,
> since that's the major way of selling lots of systems. There were
> forays into financial software, engineering, visualization, web hosting,
> information storage "curated" (Alpha). All of these were however
> premised on the sale of the system to users, or selling of online
> services (or maybe ads?)

Maybe not ads -- I had a conversation with Stephen Wolfram a few years
about ads supporting their online products, and he said they just
wouldn't bring in enough revenue to make it worth it, given the volume
of traffic they got or estimated they would get. Alpha's business
model seems to be selling subscriptions to end users and integration
with Bing/Siri/etc.

> If you insist on free and open source, it seems to me that you
> get nothing from your markets. So a killer app does nothing for you.

I insist on free and open source. Selling software licenses doesn't
work at all in that context. However, because of the shift (back) to
cloud-based software, it is now *not* true that using/creating free
and open source software means that you get nothing from your markets.
It is common to add significant value while still building a site
using open source software.

> I suggested that the market for my (free, open source) TILU online
> integration
> service could be the selling of advertising, presumably to freshman
> calculus students. ( soft drinks? acne medication? ) but never
> pursued it. Maybe online ads for Sage???

We have been experimenting with actually paying ($20/month) for
advertising for SageMathCloud. I've attached a screenshot showing
the acquisition channels for SMC during the last month, and how well
each has done. Paid advertising has by far the highest bounce rate,
but it also has the highest percentage of new users.

> Otherwise,I think you are left with finding and impressing one
> of those rich people who thinks math programs are kool.

Thanks -- please donate at

https://www.washington.edu/giving/make-a-gift/

> Good luck.

Thanks!!

[Dr. David Kirkby (Kirkby Microwave Ltd)]

On 3 October 2015 at 16:38, rjf <fat...@gmail.com> wrote:

Most software developers seeking funding need a "killer app".  I don't

know that Mathematica has one -- but maybe it is STEM education,

since that's the major way of selling lots of systems.  There were

forays into financial software, engineering, visualization, web hosting,

information storage "curated"  (Alpha).  All of these were however

premised on the sale of the system to users, or selling of online

services  (or maybe ads?)

I've often wondered where Wolfram Research make their money. Clearly Mathematica is their main product, but based on personal experience, and job adverts, I rarely see any employer wanting skills in Mathematica. Contrast that with MATLAB, and there are a huge number of jobs calling for MATLAB skills. It makes me think that the only significant purchasers of Mathematica are academic institutions.

So I don't know what their "killer app" is, as I have never seen any evidence to suggest they have one. I can only think it is selling to academia - as to what areas, I have no idea.

Dave

[Jason Grout]

We have several implementations of such a thing in Jupyter community.
Matplotlib has a basic implementation of this if you use "%matplotlib
notebook" to change to the interactive backend. Bokeh may be what you
saw, as it works also without a kernel running. We (at Bloomberg) also
just open-sourced a plotting library we've been working on at Bloomberg
for the last year or so which has interactivity, based on the Jupyter
widgets: https://github.com/bloomberg/bqplot.

Thanks,

Jason