why sage is useful for me

24 views
Skip to first unread message

Ondrej Certik

unread,
Jul 22, 2008, 6:17:13 AM7/22/08
to sage-...@googlegroups.com
Hi,

I just wanted to share why Sage is (or will be soon) useful for me.

1) I have a program that I do for my master thesis, it's some finite
elements method + electronic structure calculations and my boss gave
me access to some solaris very fast boxes,
and I told him, thanks the access works, I only need to find "a while"
to install scipy+numpy+pysparse+umfpack+netcdfg+mercurial. Michael
Abshoff told me it's easy, but I just remember I gave up installing
scipy+umfpack on some old Debian boxes in Munich, so I just installed
Sage instead (it doesn't have umfpack yet though, or it didn't back
then).
The irony is that I comaintain scipy in Debian sid and it works like a
charm in there, but when we tried to install the scipy from sid on the
old debian, it pulled a hell of dependencies including libc and gcc
upgrade (probably due to the recent gfortran transition in Debian) and
we just didn't do it, because it would mean an upgrade to sid of the
whole machine.
Anyway, so when I told my boss the packages I need to install on the
solaris to even get my program to run, he just replied "you need to
write your program in a way so that other people can actually use
it!". And he was right.

So, mabshoff -- Solaris port +1.

So when I get to it, hopefully soon, I think I'll start creating spkg
packages of missing stuff that I need and then just use the fact that
Sage just works on any system, so that other people can actually run
my programs.

So, none of the above has anything in common with symbolic
manipulation. But one of the aims of Sage is to be a viable
alternative to Matlab, so that's about it.


2) There is a nonzero chance I'll be teaching some undergrad calculus
stuff in a year or two and so I was thinking which programs (if any)
I'd use and the constrain will probably be windows. So, I would like
to use python + sympy, because I'd like to explain how to actually do
limits and integrals and derivatives on the computer (to show that
actually the way it is taught at schools, for example my school, is
old fashioned, because in practise, the computer uses different
algorithms than those that they teach us and so one should be at least
aware of this how things are done in practise), and sympy has a
working pure python implementation. But then, installing python+sympy
in windows is not convenient -- it's about 10 clicks with mouse and it
installs somewhere to C:\Python2.5 by default and I don't understand
windows much but I just get the feeling that it messes up with the
registry as well and I don't like this at all. Another problem is that
it installs, but then I need to start the shell (cmd.exe) and run
python in it (or ipython). I would like to just download one tarball,
unpack it and run it. That's it.

So, mabshoff -- Sage windows port +1.

Well, if I could even run my scipy+numpy linux programs in windows
easily using sage, that would be awesome!


So the above are rather side effects of the main Sage's aim, but those
are things that would make my life much more easier.

Ondrej

William Stein

unread,
Jul 22, 2008, 8:11:16 AM7/22/08
to sage-...@googlegroups.com

Thanks!

> So, none of the above has anything in common with symbolic
> manipulation. But one of the aims of Sage is to be a viable
> alternative to Matlab, so that's about it.

This is absolutely one of the core goals of Sage.

>
>
> 2) There is a nonzero chance I'll be teaching some undergrad calculus
> stuff in a year or two and so I was thinking which programs (if any)
> I'd use and the constrain will probably be windows. So, I would like
> to use python + sympy, because I'd like to explain how to actually do
> limits and integrals and derivatives on the computer (to show that
> actually the way it is taught at schools, for example my school, is
> old fashioned, because in practise, the computer uses different
> algorithms than those that they teach us and so one should be at least
> aware of this how things are done in practise), and sympy has a
> working pure python implementation. But then, installing python+sympy
> in windows is not convenient -- it's about 10 clicks with mouse and it
> installs somewhere to C:\Python2.5 by default and I don't understand
> windows much but I just get the feeling that it messes up with the
> registry as well and I don't like this at all. Another problem is that
> it installs, but then I need to start the shell (cmd.exe) and run
> python in it (or ipython). I would like to just download one tarball,
> unpack it and run it. That's it.

You could also just run a sage notebook server with some published worksheets.
That's what I would do in the above situation. Then the students don't have
to install anything.

> So, mabshoff -- Sage windows port +1.
>
> Well, if I could even run my scipy+numpy linux programs in windows
> easily using sage, that would be awesome!
>
>
> So the above are rather side effects of the main Sage's aim, but those
> are things that would make my life much more easier.

I think they are actually one of the three cores aims of Sage:
1. Distribution
2. New library
3. Interfaces

William

Dr. David Kirkby

unread,
Jul 22, 2008, 10:33:00 AM7/22/08
to sage-devel
On Jul 22, 11:17 am, "Ondrej Certik" <ond...@certik.cz> wrote:
> Hi,
>
> I just wanted to share why Sage is (or will be soon) useful for me.
>
> 1) I have a program that I do for my master thesis, it's some finite
> elements method + electronic structure calculations and my boss gave
> me access to some solaris very fast boxes,

Is that UltraSPARC or x86/x64 (i.e. AMD/Intel) based?

I'd be interested what you have access to.

> So, mabshoff -- Solaris port +1.

Me too wanting that.

> 2) There is a nonzero chance I'll be teaching some undergrad calculus
> stuff in a year or two and so I was thinking which programs (if any)
> I'd use and the constrain will probably be windows.

Why should it be windows?

I feel Windows has some advantages over Solaris or Linux for general
office/home use. In particular, you don't need to be as computer savvy
to do certain basic tasks, like install printers.

But I feel Unix/Linux systems are a much better platform for
scientific computing. If you use the principle of "best tool for the
job" then I doubt Windows can be considered the best tool.

Why encourage people to a system which seems to have dramatically
increased in price in recent years and gets more and more bloated?

When I bought my first PC, the hardware cost just over £2000 ($4000)
and the operating system license was about £50 ($100). Now, the cost
of the OS is a much higher percentage of the total cost.


> So, mabshoff -- Sage windows port +1.

I know he is on to that.

It's obvious that given the large number of windows systems, that it
will increase Sage's popularity more than ports to Solaris, HP-UX or
any other platform ever will. (Personally I'm more keen on Solaris
port, but objectively one would have to say a Windows one would do
more for Sage's popularity. And that is important.)

> So the above are rather side effects of the main Sage's aim, but those
> are things that would make my life much more easier.

I just wish I'd had a decent open-source alternative to Mathematica
when I started my MSc. Unfortunately I've used Mathematica on/off for
the last 17-18 years, so have a lot of time spent using it. I'd rather
not have to inflict that anyone else, so they get locked into a
system. (I think Mathematica is a good system, just too expensive and
proprietary).

mabshoff

unread,
Jul 22, 2008, 11:44:43 AM7/22/08
to sage-devel


On Jul 22, 7:33 am, "Dr. David Kirkby" <david.kir...@onetel.net>
wrote:
> On Jul 22, 11:17 am, "Ondrej Certik" <ond...@certik.cz> wrote:
>
> > Hi,

Hi Ondrej,

> > I just wanted to share why Sage is (or will be soon) useful for me.
>
> > 1) I have a program that I do for my master thesis, it's some finite
> > elements method + electronic structure calculations and my boss gave
> > me access to some solaris very fast boxes,
>
> Is that UltraSPARC or x86/x64 (i.e. AMD/Intel) based?
>
> I'd be interested what you have access to.

I could relatively painlessly build you a 64 bit self contained gcc,
binutils, shellutils, python, ATLAS, numpy 1.1.1, scipy SVN and hg if
you left me know the CPU target you want. I have to offer Sparc US
IIIi, Core2 Quad SSE or Opteron.

> > So, mabshoff -- Solaris port +1.
>
> Me too wanting that.

3.0.6 contains about five Solaris build fixes, so the list is getting
shorter. We mainly did a stabilization release, so some of the more
risky stuff was left out.

> > 2) There is a nonzero chance I'll be teaching some undergrad calculus
> > stuff in a year or two and so I was thinking which programs (if any)
> > I'd use and the constrain will probably be windows.
>
> Why should it be windows?
>
> I feel Windows has some advantages over Solaris or Linux for general
> office/home use. In particular, you don't need to be as computer savvy
> to do certain basic tasks, like install printers.
>
> But I feel Unix/Linux systems are a much better platform for
> scientific computing. If you use the principle of "best tool for the
> job" then I doubt Windows can be considered the best tool.
>
> Why encourage people to a system which seems to have dramatically
> increased in price in recent years and gets more and more bloated?
>
> When I bought my first PC, the hardware cost just over £2000 ($4000)
> and the operating system license was about £50 ($100). Now, the cost
> of the OS is a much higher percentage of the total cost.

Well, there is a vast people using Windows out there either by choice
or because they don't know about alternatives. And Sage can only
benefit from being portable. In the end it is all about bringing
mathematical OS to the people and not convincing them to switch OSes.
They will hopefully learn that Open Source is something good and high
quality, but if other people use Windows I am perfectly fine with it.

> > So, mabshoff -- Sage windows port +1.
>
> I know he is on to that.
>
> It's obvious that given the large number of windows systems, that it
> will increase Sage's popularity more than ports to Solaris, HP-UX or
> any other platform ever will. (Personally I'm more keen on Solaris
> port, but objectively one would have to say a Windows one would do
> more for Sage's popularity. And that is important.)
>
> > So the above are rather side effects of the main Sage's aim, but those
> > are things that would make my life much more easier.
>
> I just wish I'd had a decent open-source alternative to Mathematica
> when I started my MSc. Unfortunately I've used Mathematica on/off for
> the last 17-18 years, so have a lot of time spent using it. I'd rather
> not have to inflict that anyone else, so they get locked into a
> system. (I think Mathematica is a good system, just too expensive and
> proprietary).

:)

Cheers,

Michael

Dr. David Kirkby

unread,
Jul 22, 2008, 2:51:39 PM7/22/08
to sage-devel


On Jul 22, 4:44 pm, mabshoff <mabsh...@googlemail.com> wrote:

> Well, there is a vast people using Windows out there either by choice
> or because they don't know about alternatives.

True, but I suspect the first year students under discussion should be
aware of other systems such as Linux. They would have the choice to
use Linux if they wanted. I can't believe it would be hard to find a
Linux box in a university now if you wanted to use one. Even easier to
install it on their own PC.

> And Sage can only
> benefit from being portable.

Agreed - a Windows port is essential. More so than Solaris.

> In the end it is all about bringing
> mathematical OS to the people and not convincing them to switch OSes.

I do not agree with you there!

If Linux is a more suitable platform for scientific computing than
Windows, it would be wise to encourage first year students to use
Linux for scientific computing. Unlike staff working for most large
corporations, students will not be tied to rigid corporate policies.
They can install software on their own personal machines.

I've spent the last couple of weeks doing "scientific computing" and I
believe it would have been considerably more difficult under Windows
than it was under Solaris.

I've also spent a few days writing a scientific report under Solaris.
I'm the first to admit, that would have been no more difficult, and
perhaps even sligthly easier under Windows, especially as I had to use
files Visio format files sent to me. I had to get a colleague to put
them in a usable format. He could not use postscript (why I will never
no), so I had to make do with PDF. Not ideal. But it would have been
easier for me to just use Windows for that part of it.

So I'm not proposing a blanket "convince everyone to use Linux and
scrap Windows", but I do believe if Linux/Unix is a better platform
for scientific computing, then it is wise to encourage first year
student to use Linux/Unix.

> They will hopefully learn that Open Source is something good and high
> quality, but if other people use Windows I am perfectly fine with it.

I'm happy with it, but feel its unwise to enough it use for scientific
computing to a group of individuals who have no compelling reasons
(ie. corporate policy) to use Windows for scientific computing.

> Michael

Dave

Andrey Novoseltsev

unread,
Jul 22, 2008, 4:26:25 PM7/22/08
to sage-devel
> > Well, there is a vast people using Windows out there either by choice
> > or because they don't know about alternatives.
>
> True, but I suspect the first year students under discussion should be
> aware of other systems such as Linux. They would have the choice to
> use Linux if they wanted. I can't believe it would be hard to find a
> Linux box in a university now if you wanted to use one. Even easier to
> install it on their own PC.
>
> > In the end it is all about bringing
> > mathematical OS to the people and not convincing them to switch OSes.
>
> I do not agree with you there!
>
> If Linux is a more suitable platform for scientific computing than
> Windows, it would be wise to encourage first year students to use
> Linux for scientific computing. Unlike staff working for most large
> corporations, students will not be tied to rigid corporate policies.
> They can install software on their own personal machines.

I think sticking with Windows or trying Linux is a question of
audience. Most calculus students probably don't do a lot of scientific
computing and will not spend too much time on a particular Calculus
class, in which case it is quite inconvenient to install another
operating system for computing limits for a number of reasons.

First, the installation is long and probably requires repartitioning
of the hard disk. And it is preferably not to loose any data. I got my
laptop with Windows and a single NTFS partition and it took me a while
to find out a way to decrease the size of this partition without
destroying data (I think some SUSE distribution was able to do it, but
not two or three others which I have tried first).

Second, for everyday home use linux maybe less convenient, e.g. I was
not able to easily figure out a way to play DVDs, they didn't work by
default. And not many popular games will work under linux, if any,
which is probably important for many students.

Third, even if you have two OS installed on your computer, it is
annoying to switch between them if it requires rebooting.

Due to all the reasons above, I REALLY like VMware image of sage to
run locally and after starting using it I removed the Linux partition
from my computer and made just a separate disk for the image files.
But my favorite way is to run sage remotely and use only browser on my
machine. And I think that this is also the way many calculus students
will like most, even when there will be a native sage port - no need
to install anything on your machine and ability to continue your work
from any computer in the library, under any operating system!

So I like both linux and windows and the fact that sage notebook runs
in the same way on any system for the same zero price, which in my
opinion is also one of the big advantages over other systems. I have
Maple on my computer, under Windows, but if I want to use it also
under linux I need to buy second licence. Even with academic prices it
is expensive.

root

unread,
Jul 22, 2008, 6:10:53 PM7/22/08
to sage-...@googlegroups.com, sage-...@googlegroups.com
> > Well, there is a vast people using Windows out there either by choice
> > or because they don't know about alternatives.
>
> True, but I suspect the first year students under discussion should be
> aware of other systems such as Linux. They would have the choice to
> use Linux if they wanted. I can't believe it would be hard to find a
> Linux box in a university now if you wanted to use one. Even easier to
> install it on their own PC.

Just as a data point: While I was at City College in New York I
started a project called Doyen to make a Live CD as a computer
algebra platform. I advertised in the computer science department
for 2 students with a knowledge of Linux for a paid position.
Ten students showed up for interviews, all of them were computer
science majors and most of them were seniors. Eight of the ten
had not heard of Linux.

This was 4 years ago. Things might have changed. Your school may vary.

Tim


William Stein

unread,
Jul 22, 2008, 6:17:06 PM7/22/08
to sage-...@googlegroups.com

+1

I'll go so far to say that it's perfectly fine with me for somebody to
*like* to
use Windows. In fact, I used Windows by choice
on my *laptop* for a full year in 2003, and some of the time I really
really liked it. I would have laughed if somebody told me that
they had made a viable alternative to the Ma's but that
if I wanted to use it I had to switch from Windows back to Linux.

> Third, even if you have two OS installed on your computer, it is
> annoying to switch between them if it requires rebooting.

I agree -- it is MASSIVELY annoying, and completely unreasonable
to ask of users to reboot to use your program. Especially
given that a lot of people like to stay
logged into chat programs, etc. This is also a strong argument
against live boot cd's, unless they are run via virtualization (e.g., vmware).

> Due to all the reasons above, I REALLY like VMware image of sage to
> run locally and after starting using it I removed the Linux partition
> from my computer and made just a separate disk for the image files.
> But my favorite way is to run sage remotely and use only browser on my
> machine. And I think that this is also the way many calculus students
> will like most, even when there will be a native sage port - no need
> to install anything on your machine and ability to continue your work
> from any computer in the library, under any operating system!

Yep. That really agrees with how I think tons of people feel.
I think it's closely related to something Google and numerous
other companies noticed a couple of years ago.

Ondrej Certik

unread,
Jul 22, 2008, 7:20:54 PM7/22/08
to sage-...@googlegroups.com
On Tue, Jul 22, 2008 at 4:33 PM, Dr. David Kirkby
<david....@onetel.net> wrote:
>
> On Jul 22, 11:17 am, "Ondrej Certik" <ond...@certik.cz> wrote:
>> Hi,
>>
>> I just wanted to share why Sage is (or will be soon) useful for me.
>>
>> 1) I have a program that I do for my master thesis, it's some finite
>> elements method + electronic structure calculations and my boss gave
>> me access to some solaris very fast boxes,
>
> Is that UltraSPARC or x86/x64 (i.e. AMD/Intel) based?
>
> I'd be interested what you have access to.

It's AMD opteron. I thought they run solaris, but actually I've found
out they run opensuse. So I tried to install Sage, because that should
work, but to my big surprise it doesn't compile:

http://groups.google.com/group/sage-support/msg/89d14e6cb5eaf182

but it should be easier to fix than the solaris port. Which is a good
news to me, maybe I'll be able to use the boxes soon.

> Why should it be windows?

Andrey, root and William has already said the points. I can only add
my own point of view, I am right wing thinking, so I don't try and
believe in persuadig people to think differently (or the same as I
do), I rather try to understand why people think what they think and
start from that. And try to do my actions in a way that if people do
what is good for them, it's also good for me.

So if the students use windows, they have reasons for that. And
because switching is pain, it's better for me to use the fact that
they use Windows as an axiom and make my program run on windows.
Because then they'll try it. Otherwise they won't.

If it comes to my own laptop, I of course use Debian, because it's the
best. :) I believe that everyone is using the system that is the best
for him. By (my) definition.

Ondrej

mhampton

unread,
Jul 23, 2008, 5:51:47 AM7/23/08
to sage-devel

> > Due to all the reasons above, I REALLY like VMware image of sage to
> > run locally and after starting using it I removed the Linux partition
> > from my computer and made just a separate disk for the image files.
> > But my favorite way is to run sage remotely and use only browser on my
> > machine. And I think that this is also the way many calculus students
> > will like most, even when there will be a native sage port - no need
> > to install anything on your machine and ability to continue your work
> > from any computer in the library, under any operating system!
>
> Yep. That really agrees with how I think tons of people feel.
> I think it's closely related to something Google and numerous
> other companies noticed a couple of years ago.

I can second this, and I have taught using Sage. Students really like
a remote server (or two) that they can just log in to from anywhere,
that auto-saves their work. It makes a number of things easier for
the instructor as well, since everything is centralized.

-M. Hampton

Ondrej Certik

unread,
Jul 23, 2008, 6:30:45 AM7/23/08
to sage-...@googlegroups.com

That sounds like the way to go. BTW I think it can be greatly
improved, for example I think the login should not be required if one
doesn't
need to have things saved. Something like:

http://live.sympy.org/

Where anyone can immediatelly start typing commands and use it.
However, the interface needs to be greatly improved, this is just a
proof of concept, but it works. I think it's like a pyramid. You start
with something like live.sympy.org where you don't need any login,
nothing, you just use it and it works. Then you decide you'd like to
save your state. So you create a login and save your worksheets. Then
you decide you'd actually run it on your computer, either for greater
speed, or to use data from your computer or to have a better control
of it. You run it on windows for example. And then you decide you'd
actually do a lot of development, and you might switch to linux or Mac
Os X. Or you may install the dev tools in windows.

Ondrej

Harald Schilly

unread,
Jul 23, 2008, 8:32:00 AM7/23/08
to sage-devel
On 23 Jul., 12:30, "Ondrej Certik" <ond...@certik.cz> wrote:
>
> I think it's like a pyramid. You start
> with something like live.sympy.org where you don't need any login...

Yes, I think the same way. I know there are a lot of crazy scripts out
there, which try to mess up with every form they can submit, so maybe,
live.sympy.org is too easy to access and a captcha-type entry page
storing a one-time code as a cookie is necessary. But yes, the entry
barriers need to be low. The same applies for every further step - the
transition steps should be easy, but some are already (think of python
everywhere and so on)

Also, a year ahead, think of tools like google gears or a bit later
HTML 5. Then, it is possible to store all the worksheet data locally
inside the users browser in a sql-lite database. So, the data could be
tied to the users local computer account, the "computational power" is
in the network; or by changing the host, locally on the machine. (I
think this is currently impossible due to cross-site restrictions, but
maybe there will be a way to allow this kind of storage (box pops up
an asks) - at least worth to request in the gears group ;) [and
sharing worksheets would mean to transfer the data to a public server
or something like that...]

harald

William Stein

unread,
Jul 23, 2008, 8:46:52 AM7/23/08
to sage-...@googlegroups.com
On Wed, Jul 23, 2008 at 2:32 PM, Harald Schilly
<harald....@gmail.com> wrote:
>
> On 23 Jul., 12:30, "Ondrej Certik" <ond...@certik.cz> wrote:
>>
>> I think it's like a pyramid. You start
>> with something like live.sympy.org where you don't need any login...
>
> Yes, I think the same way. I know there are a lot of crazy scripts out
> there, which try to mess up with every form they can submit, so maybe,
> live.sympy.org is too easy to access

Sage is one such crazy script :-)

sage: magma_free('Factorization(100000000000000000000090130000000000000000000117)')
[ <10000000000000000000009, 1>, <10000000000000000000000013, 1> ]


> and a captcha-type entry page
> storing a one-time code as a cookie is necessary. But yes, the entry
> barriers need to be low. The same applies for every further step - the
> transition steps should be easy, but some are already (think of python
> everywhere and so on)

+1; let's do it...

> Also, a year ahead, think of tools like google gears or a bit later
> HTML 5. Then, it is possible to store all the worksheet data locally
> inside the users browser in a sql-lite database. So, the data could be
> tied to the users local computer account, the "computational power" is
> in the network; or by changing the host, locally on the machine. (I
> think this is currently impossible due to cross-site restrictions, but
> maybe there will be a way to allow this kind of storage (box pops up
> an asks) - at least worth to request in the gears group ;) [and
> sharing worksheets would mean to transfer the data to a public server
> or something like that...]

Cool! Let's make it happen!

William

Timothy Clemans

unread,
Jul 23, 2008, 12:16:43 PM7/23/08
to sage-...@googlegroups.com
I actually have TryPy in my home directory
http://sage.math.washington.edu/home/tclemans/trypy/ I think it's a
better Python AJAX console.

Dr. David Kirkby

unread,
Jul 23, 2008, 12:27:34 PM7/23/08
to sage-devel


On 22 Jul, 23:10, root <d...@axiom-developer.org> wrote:

> Just as a data point: While I was at City College in New York I
> started a project called Doyen to make a Live CD as a computer
> algebra platform. I advertised in the computer science department
> for 2 students with a knowledge of Linux for a paid position.
> Ten students showed up for interviews, all of them were computer
> science majors and most of them were seniors. Eight of the ten
> had not heard of Linux.
>
> This was 4 years ago. Things might have changed. Your school may vary.
>
> Tim

You really surprise me there.

I don't work in a uni any more, but the last time I did (about 4 years
ago), a very large number of people were running linux. And we were
not a computer science department (medical physics). Whereas Solaris
systems were offically suppored and maintained by the IT staff, Linux
was normally maintained by the individual who run it, or perhaps one
person in a group.

At the time I left, there was serious talk of running a lot of linux
boxes and having them administered by the IT guys, and individuals not
having root access.

In the UK there are numerous magazines on the newsagents shelves
devoted exclusively to linux, and I think every computer magazine will
have a linux column. I don't think the situation was much different 4
years ago.

I know several people who have removed Vista from their machines, but
most of them install XP, not Linux. People tend to install linux
because they want to, rather than because they are unhappy with
Vista.

I wonder if this varies much by country?





Harald Schilly

unread,
Jul 23, 2008, 1:32:24 PM7/23/08
to sage-devel
On Jul 23, 6:27 pm, "Dr. David Kirkby" <david.kir...@onetel.net>
wrote:
>
> I wonder if this varies much by country?

Yes it does. Depends on the income of people (e.g. in some countries
windows is not affordable) and the perception of windows. Here in
europe, there is a feeling that buying ms-licenses of any sort pours
money over the ocean. It would be better to invest into european based
technology. That's one of the reasons, why there are efforts to
migrate federal infrastructure to linux. Example, here at my
university you can get linux on your desktop, administrated by the
central it department (its an old version and there are troubles, but
at least a sign into that direction)

And something else about dual booting: don't forget the last years
prices. I think, those who like computers of any kind already have 2
or more. Just encourage them to install linux on their "old" laptop/pc
and then they have one for games and one for work and no need for dual
booting...

h

Simon King

unread,
Jul 23, 2008, 1:49:02 PM7/23/08
to sage-devel
On Jul 23, 7:32 pm, Harald Schilly <harald.schi...@gmail.com> wrote:
> > I wonder if this varies much by country?
> Yes it does.

I was at universities in Strasbourg and various places in Germany.
There, it was actually *difficult* to get Windows on your desk.
Essentially the only people ever using Windows were the secretaries.

Visiting Chicago, i found for the first time a mathematician using a
Mac in her office.

And at some conference in Ann Arbor i actually got quite angry: The
computers for participants only had Windows, but they did *not* have
ssh (not even telnet). I was feeling like in a desert.

So, indeed, it depends on the country. The US seem to be particularly
fond of Windows.

Yours
Simon

Ondrej Certik

unread,
Jul 23, 2008, 4:26:21 PM7/23/08
to sage-...@googlegroups.com
On Wed, Jul 23, 2008 at 2:46 PM, William Stein <wst...@gmail.com> wrote:
>
> On Wed, Jul 23, 2008 at 2:32 PM, Harald Schilly
> <harald....@gmail.com> wrote:
>>
>> On 23 Jul., 12:30, "Ondrej Certik" <ond...@certik.cz> wrote:
>>>
>>> I think it's like a pyramid. You start
>>> with something like live.sympy.org where you don't need any login...
>>
>> Yes, I think the same way. I know there are a lot of crazy scripts out
>> there, which try to mess up with every form they can submit, so maybe,
>> live.sympy.org is too easy to access
>
> Sage is one such crazy script :-)
>
> sage: magma_free('Factorization(100000000000000000000090130000000000000000000117)')
> [ <10000000000000000000009, 1>, <10000000000000000000000013, 1> ]

Google fixes that by having 10s limit, then it kills the code.

Let's do it the opposite way. Let people access it completely freely,
and fix problems (i.e. add restrictions) only if needed.

With sympy it's very easy, because it's pure python, so I just leave
all the nasty scripts to google to handle. :)

>> and a captcha-type entry page
>> storing a one-time code as a cookie is necessary. But yes, the entry
>> barriers need to be low. The same applies for every further step - the
>> transition steps should be easy, but some are already (think of python
>> everywhere and so on)
>
> +1; let's do it...
>
>> Also, a year ahead, think of tools like google gears or a bit later
>> HTML 5. Then, it is possible to store all the worksheet data locally
>> inside the users browser in a sql-lite database. So, the data could be
>> tied to the users local computer account, the "computational power" is
>> in the network; or by changing the host, locally on the machine. (I
>> think this is currently impossible due to cross-site restrictions, but
>> maybe there will be a way to allow this kind of storage (box pops up
>> an asks) - at least worth to request in the gears group ;) [and
>> sharing worksheets would mean to transfer the data to a public server
>> or something like that...]
>
> Cool! Let's make it happen!


Yes, let's do it. I am not good at web stuff though, so I'd prefer if
someone else did it. :)

Ondrej

Ondrej Certik

unread,
Jul 23, 2008, 4:36:39 PM7/23/08
to sage-...@googlegroups.com
On Wed, Jul 23, 2008 at 6:16 PM, Timothy Clemans
<timothy...@gmail.com> wrote:
>
> I actually have TryPy in my home directory
> http://sage.math.washington.edu/home/tclemans/trypy/ I think it's a
> better Python AJAX console.

Could you please put it to the google app engine, so that we can play
with it? I'd spare all of us time to install it. :)

Ondrej

Harald Schilly

unread,
Jul 23, 2008, 4:50:26 PM7/23/08
to sage-devel
On Jul 23, 10:26 pm, "Ondrej Certik" <ond...@certik.cz> wrote:
> >> Also, a year ahead, think of tools like google gears or a bit later
> >> HTML 5....
>
> > Cool!  Let's make it happen!
>
> Yes, let's do it.  I am not good at web stuff though, so I'd prefer if
> someone else did it. :)
>

ok, good, i want to do something with gears anyway, too. let's make it
this way: i read the api and groups, think about what could be done,
write a "paper" and that's then the plan to implement it.

h

Dr. David Kirkby

unread,
Jul 23, 2008, 6:11:19 PM7/23/08
to sage-devel


On Jul 23, 6:49 pm, Simon King <k...@mathematik.uni-jena.de> wrote:
> On Jul 23, 7:32 pm, Harald Schilly <harald.schi...@gmail.com> wrote:
>
> > > I wonder if this varies much by country?
> > Yes it does.
>
> I was at universities in Strasbourg and various places in Germany.
> There, it was actually *difficult* to get Windows on your desk.
> Essentially the only people  ever using Windows were the secretaries.

Linux is definately popular in Germany. I've found this working with a
uni and a commercial company in Germany recently. Both make a lot of
use of Linux, but they use Windows too, so I can't say it would have
been difficult to find a Windows machine.

I get a few interested looks from people if they see my laptop runs
Solaris, but most don't even notice - even on the odd occasions I loan
it to someone so they can look something up on a web site.

It must be said, Solaris on a laptop is not common in the UK.

> Visiting Chicago, i found for the first time a mathematician using a
> Mac in her office.

> And at some conference in Ann Arbor i actually got quite angry: The
> computers for participants only had Windows, but they did *not* have
> ssh (not even telnet). I was feeling like in a desert.

I guess the conference organisers would argue it for security reasons.
But I would agree it rather makes things difficult.

One option is to put putty on a USB memory stick and configure your
router to redirect port 443 or some other to port 22. Then you can use
SSH, even if they block port 22 going out. Once you get to one
machine, you can use that to go elsewhere.

> So, indeed, it depends on the country. The US seem to be particularly
> fond of Windows.

I guess that is no surprise.

What I find so surprising here is the number of people removing Vista
from their machines and putting XP on. It's almost an epidemic. A
couple of weeks back I was working in Scotland with 5 other engineers.
Not a single engineer had Vista on his laptop (I had Solaris, the rest
XP). Yet I know at least two of those laptops were bought with Vista.

I have never before given much thought to the geographical
distribution of operating systems, but it seems there certainly are
differences between the US and Europe.

Harald Schilly

unread,
Jul 23, 2008, 7:46:14 PM7/23/08
to sage-devel
On Jul 24, 12:11 am, "Dr. David Kirkby" <david.kir...@onetel.net>
wrote:
>
> I have never before given much thought to the geographical
> distribution of operating systems, but it seems there certainly are
> differences between the US and Europe.

Well, I have some (as always incomplete) data from sagemath.org for
this month July ... [OS visits %]

USA / Operating Systems:
Windows 3,551 47.08%
Macintosh 2,052 27.21%
Linux 1,871 24.81%

Western Europe (France, Germany, Austria, not Spain, not Italy)
Linux 1,793 43.08%
Windows 1,487 35.73%
Macintosh 815 19.58%

Northern Europe (UK, Finland, Sweden ...)
Windows 993 52.73%
Linux 616 32.71%
Macintosh 245 13.01%

harald

Robert Bradshaw

unread,
Jul 24, 2008, 12:22:18 AM7/24/08
to sage-...@googlegroups.com

Western Europe < USA < Northern Europe (for Windows) is really
interesting--I'd have expected USA to have the highest percentage of
Windows visits (but not by much) There's bound to be a huge bias for
sagemath.org one way or another so I am taking this with a huge grain
of salt.

In terms of Tim Daly's experience of CS majors that haven't even
heard of Linux, all I can say is Wow! My experience has been (even
back 4 years ago) that nearly everyone that worked/studied in the CS
departments knew a bit if Linux, even if many didn't choose it as
their primary OS.

- Robert

William Stein

unread,
Jul 24, 2008, 5:27:17 AM7/24/08
to sage-...@googlegroups.com
On Wed, Jul 23, 2008 at 10:26 PM, Ondrej Certik <ond...@certik.cz> wrote:
>
> On Wed, Jul 23, 2008 at 2:46 PM, William Stein <wst...@gmail.com> wrote:
>>
>> On Wed, Jul 23, 2008 at 2:32 PM, Harald Schilly
>> <harald....@gmail.com> wrote:
>>>
>>> On 23 Jul., 12:30, "Ondrej Certik" <ond...@certik.cz> wrote:
>>>>
>>>> I think it's like a pyramid. You start
>>>> with something like live.sympy.org where you don't need any login...
>>>
>>> Yes, I think the same way. I know there are a lot of crazy scripts out
>>> there, which try to mess up with every form they can submit, so maybe,
>>> live.sympy.org is too easy to access
>>
>> Sage is one such crazy script :-)
>>
>> sage: magma_free('Factorization(100000000000000000000090130000000000000000000117)')
>> [ <10000000000000000000009, 1>, <10000000000000000000000013, 1> ]
>
> Google fixes that by having 10s limit, then it kills the code.

I actually wrote (6-7 years ago)the web backend that is being
scripted/scraped by magma_free above, and I limited computation to
30s.

> Let's do it the opposite way. Let people access it completely freely,
> and fix problems (i.e. add restrictions) only if needed.
>
> With sympy it's very easy, because it's pure python, so I just leave
> all the nasty scripts to google to handle. :)
>
>>> and a captcha-type entry page
>>> storing a one-time code as a cookie is necessary. But yes, the entry
>>> barriers need to be low. The same applies for every further step - the
>>> transition steps should be easy, but some are already (think of python
>>> everywhere and so on)
>>
>> +1; let's do it...
>>
>>> Also, a year ahead, think of tools like google gears or a bit later
>>> HTML 5. Then, it is possible to store all the worksheet data locally
>>> inside the users browser in a sql-lite database. So, the data could be
>>> tied to the users local computer account, the "computational power" is
>>> in the network; or by changing the host, locally on the machine. (I
>>> think this is currently impossible due to cross-site restrictions, but
>>> maybe there will be a way to allow this kind of storage (box pops up
>>> an asks) - at least worth to request in the gears group ;) [and
>>> sharing worksheets would mean to transfer the data to a public server
>>> or something like that...]
>>
>> Cool! Let's make it happen!
>
>
> Yes, let's do it. I am not good at web stuff though, so I'd prefer if
> someone else did it. :)

Same here :-)

William

Harald Schilly

unread,
Jul 24, 2008, 5:42:29 AM7/24/08
to sage-devel
On Jul 24, 6:22 am, Robert Bradshaw <rober...@math.washington.edu>
wrote:
> Western Europe < USA < Northern Europe (for Windows) is really  
> interesting--I'd have expected USA to have the highest percentage of  
> Windows visits ...

yes, also, those numbers have higher error margins if there is less
data. but some details:

UK
Windows 668 59.75%
Linux 271 24.24%
Macintosh 156 13.95%
SunOS 16 1.43%
(including Sun, because it is above 1% .. that's not everywhere the
case!)
(and think about that: the european "center" of microsoft is ireland,
part of UK, because of the cheap taxes - so it seems like geography
does matter ...)

Seattle, WA, USA
Macintosh 174 44.16%
Windows 158 40.10%
Linux 62 15.74%

whole WA, USA
Windows 261 44.85%
Macintosh 220 37.80%
Linux 101 17.35%

Germany
Linux 931 45.93%
Windows 707 34.88%
Macintosh 351 17.32%

also keep in mind, that his is not representative. just a snapshot of
those "early adopters" who are interested in sage. that's not the real
picture.

h

John Cremona

unread,
Jul 24, 2008, 3:53:57 PM7/24/08
to sage-...@googlegroups.com
I'm still recovering from learning that I live in "Northern Europe"
and not "Western Europe". And now you say that "Ireland is part of
the UK"!

Until the early 20th century Ireland was part of Great Britain (the
British Empire). Then it was divided: the catholic south became
Eire, the Republic of Ireland, while the protestant north became
Northern Ireland which is part of the UK (= "United Kingdom of Great
Britain and Northern Ireland"). Much blood has been spilled over this
since.

Both of these two countries are members of the EU. The RoI is in the
euro zone while the UK is not. Each is independent and sets its own
taxes. I have no idea if taxes there are less than here (=UK).

I have no political axe to grind here, seriously!

John

2008/7/24 Harald Schilly <harald....@gmail.com>:

Ondrej Certik

unread,
Jul 24, 2008, 4:16:28 PM7/24/08
to sage-...@googlegroups.com
On Thu, Jul 24, 2008 at 9:53 PM, John Cremona <john.c...@gmail.com> wrote:
>
> I'm still recovering from learning that I live in "Northern Europe"
> and not "Western Europe". And now you say that "Ireland is part of
> the UK"!
>
> Until the early 20th century Ireland was part of Great Britain (the
> British Empire). Then it was divided: the catholic south became
> Eire, the Republic of Ireland, while the protestant north became
> Northern Ireland which is part of the UK (= "United Kingdom of Great
> Britain and Northern Ireland"). Much blood has been spilled over this
> since.
>
> Both of these two countries are members of the EU. The RoI is in the
> euro zone while the UK is not. Each is independent and sets its own
> taxes. I have no idea if taxes there are less than here (=UK).
>
> I have no political axe to grind here, seriously!


Haha, that made me laugh. It's like some people think that the Czech
Republic is in eastern europe, which is obviously not true.

http://en.wikipedia.org/wiki/Czech_Republic

Well, but if England can be in the Northen Eeurope, why not then.
Anyway, back to work.

Ondrej

Harald Schilly

unread,
Jul 24, 2008, 4:43:26 PM7/24/08
to sage-devel
On Jul 24, 9:53 pm, "John Cremona" <john.crem...@gmail.com> wrote:
> I'm still recovering from learning that I live in "Northern Europe"
> and not "Western Europe".  And now you say that "Ireland is part of
> the UK"!

just for the record, 1. i was tired and 2. just when i sent the mail i
thought that i should have double checked what i said.

because yes, the google statistic stuff says it is northern europe,
together with sweden and norway, but not finland (that's already
eastern europe, together with russia and czech)
and second and most important, i was wrong, down there on the list is
ireland, so it's not part of uk.
my fault
please don't spill blood on me

Ireland / Operating System
Windows 49 64.47%
Linux 17 22.37%
Macintosh 10 13.16%

at least i wasn't wrong with that analysis ;)

in the future there will be no sub-european continents. i promise.

harald

Dan Drake

unread,
Jul 24, 2008, 8:31:30 PM7/24/08
to sage-...@googlegroups.com
On Thu, 24 Jul 2008 at 08:53PM +0100, John Cremona wrote:
> I'm still recovering from learning that I live in "Northern Europe"
> and not "Western Europe". And now you say that "Ireland is part of
> the UK"!
>
> Until the early 20th century Ireland was part of Great Britain (the
> British Empire). Then it was divided: the catholic south became
> Eire, the Republic of Ireland, while the protestant north became
> Northern Ireland which is part of the UK (= "United Kingdom of Great
> Britain and Northern Ireland"). Much blood has been spilled over this
> since.
>
> Both of these two countries are members of the EU. The RoI is in the
> euro zone while the UK is not. Each is independent and sets its own
> taxes. I have no idea if taxes there are less than here (=UK).
>
> I have no political axe to grind here, seriously!

We're veering ridiculously far offtopic here, but...for anyone who gets
confused at the differences between "Britain", "the UK", "British Isles"
and so on, here's a useful Venn diagram:

http://qntm.org/?uk

Dan

--
--- Dan Drake <dr...@mathsci.kaist.ac.kr>
----- KAIST Department of Mathematical Sciences
------- http://math.kaist.ac.kr/~drake

signature.asc

David Harvey

unread,
Jul 24, 2008, 9:47:44 PM7/24/08
to sage-...@googlegroups.com

On Jul 24, 2008, at 8:31 PM, Dan Drake wrote:

> We're veering ridiculously far offtopic here, but...for anyone who
> gets
> confused at the differences between "Britain", "the UK", "British
> Isles"
> and so on, here's a useful Venn diagram:
>
> http://qntm.org/?uk

THANK YOU!!!!!

david

Ondrej Certik

unread,
Jul 27, 2008, 11:27:32 AM7/27/08
to sage-...@googlegroups.com
On Wed, Jul 23, 2008 at 1:20 AM, Ondrej Certik <ond...@certik.cz> wrote:
> On Tue, Jul 22, 2008 at 4:33 PM, Dr. David Kirkby
> <david....@onetel.net> wrote:
>>
>> On Jul 22, 11:17 am, "Ondrej Certik" <ond...@certik.cz> wrote:
>>> Hi,
>>>
>>> I just wanted to share why Sage is (or will be soon) useful for me.
>>>
>>> 1) I have a program that I do for my master thesis, it's some finite
>>> elements method + electronic structure calculations and my boss gave
>>> me access to some solaris very fast boxes,
>>
>> Is that UltraSPARC or x86/x64 (i.e. AMD/Intel) based?
>>
>> I'd be interested what you have access to.
>
> It's AMD opteron. I thought they run solaris, but actually I've found
> out they run opensuse. So I tried to install Sage, because that should
> work, but to my big surprise it doesn't compile:
>
> http://groups.google.com/group/sage-support/msg/89d14e6cb5eaf182
>
> but it should be easier to fix than the solaris port. Which is a good
> news to me, maybe I'll be able to use the boxes soon.
>
>> Why should it be windows?
>
> Andrey, root and William has already said the points. I can only add
> my own point of view, I am right wing thinking, so I don't try and
> believe in persuadig people to think differently (or the same as I

I just learned that this view is rather Libertarian:

http://en.wikipedia.org/wiki/Libertarian

Anyway.

Ondrej

Reply all
Reply to author
Forward
0 new messages