How popular is Mathematica compared to Sage? 1.71:1 is one guess.

[Dr. David Kirkby]

If you believe what Wolfram Research say, there are several million Mathematica
users. See for example

http://www.wolfram-media.com/products/mathematicabook.html

where it says about a book:

"The definitive reference and tutorial for several million enthusiastic
Mathematica users around the world"

There are several things that make me question the truth of this. One,
admittedly not very sophisticated metric, suggests the number of *active*
Mathematica users might be only 1.71 that of Sage!

First, why do I not believe this several million number?

1) As an engineer, working in several companies over the years, I've seen very
little usage of it. Plenty in academia, but little outside.

2) Do a job search on monster.com and see how many jobs require Mathematica
knowledge. Then compare it to MATLAB, and you will find far more companies want
MATLAB skills than want Mathematica. If there were several million users, I
would expect to be able to find lots of jobs requiring Mathematica skills.

3) WRI come out with a lot of ****, like for example that publishing a
'Demonstration' on

http://demonstrations.wolfram.com/index.html

counts as an academic publication! See the FAQ at

http://demonstrations.wolfram.com/FAQ.html

where it says:

---------------------------------------------------
Question. Do Demonstrations count as academic publications?

Answer. Yes. Every Demonstration undergoes a rigorous review process that checks
for quality, clarity, and accuracy, so you can count them as academic publications.

---------------------------------------------------

It sure would be easy to get a lot of publications and a chair if one could
write papers as simple as this demo!

http://demonstrations.wolfram.com/SineAndCosineGraphGenerator/

So how can one judge the popularity of Sage vs Mathematica? Well, given they
both have one main public support forum:

Sage - sage-s...@googlegroups.com
Mathematica - comp.soft-sys.math.mathematica

a comparison of the number of recent posts to these forums might give us a clue.
(Can anyone find a better way? Is this a totally flawed metric?).

As a test, I logged into Google groups about 10 days ago and quickly looked at
both the sage-s...@googlegroups.com and comp.soft-sys.math.mathematica lists.
When you do this, Google set the time of "Last Visit" to zero and the number of
"New Items" back to zero. Some time later, you can see when you last looked, and
how many posts there have been since.

Here's a screen shot.

http://sage.math.washington.edu/home/kirkby/Mathematica-vs-Sage/mathematica-vs-sage-support-requests-19-11-2009.png

The number of new posts in this period, which was about 10 days, but I can't be
sure exactly, are:

comp.soft-sys.math.mathematica 194
sage-s...@googlegroups.com 113

That's a Mathematica/Sage ratio of 1.71

That's not a huge difference. If Mathematica has several million enthusiastic
users, then Sage should have more than one million, which I very much doubt!

There are clearly problems with a direct comparison like this. It's clearly not
a rigorous statistic.

* Mathematica users can email sup...@wolfram.com, but with a turnaround time of
a few days, it is not very useful for quick questions/answers.

* Since comp.soft-sys.math.mathematica is moderated, some people do ask their
questions in sci.math.symoblic too. But there are not that many Mathematica
questions there, so I do not believe that distorts the number a lot.

* Perhaps Mathematica is so easy to use, few users need support. In contrast
Sage is so hard to use, that a higher percentage of users need to ask for
support. Personally I do not believe this to be so.

PS, you can't make similar assumptions about Solaris/HP-UX/AIX as all of them
have other major public supports forums. But Sage and Mathematica do not, so a
direct comparison of new posts, while not perfect, gives us at least a clue.

It will be interesting to see how that ratio changes over time. I might set up
another google account, register both groups, but never read them. That will
allow us see how these numbers changes over time.

Dave

[Harald Schilly]

Hi, about the publications ... big lol!

better numbers:
http://groups.google.com/group/sage-support/about > Archive table
http://groups.google.com/group/comp.soft-sys.math.mathematica/about >
Archive table

Homework: Create nice plots ;)

H

[William Stein]

Wow. LOL.

> ---------------------------------------------------
>
> It sure would be easy to get a lot of publications and a chair if one could
> write papers as simple as this demo!
>
> http://demonstrations.wolfram.com/SineAndCosineGraphGenerator/
>
> So how can one judge the popularity of Sage vs Mathematica? Well, given they
> both have one main public support forum:
>
> Sage - sage-s...@googlegroups.com
> Mathematica - comp.soft-sys.math.mathematica
>
> a comparison of the number of recent posts to these forums might give us a clue.
> (Can anyone find a better way? Is this a totally flawed metric?).

Here's another metric that I like, which measures number of users. At

http://www.wolfram.com/company/background.html

they say "With a tightly knit core of fewer than 500 employees ....".
So, let's assume

1. 500 employees
2. Each employee costs WRI $150K/year on average (including benefits, etc.)
3. An average Mathematica user pays $100/license. This includes
site licenses, which
might cover 100 active users (say, at a uni) and cost $10K, say.

That gives about 750,000 active users:

500*150000/100 = 750000

This is not so far of from Wolfram's claims.

Last time I did this calculation (in mid 2007), the Wolfram page
listed 550 employees... and when I talked to Eric Weisstein in January
2008 at the AMS meeting he said they had been hiring a lot of new
people. Given the current statement of "less than 500 employees", it
looks like they may have had some significant layoffs, due to the
financial climate in 2008-2009.

By the way, there are I think about 1500 fulltime employees at the
company that makes MATLAB.

-- William

[Jason Grout]

Okay:

http://alpha.sagenb.org/home/pub/3/

Jason


--
Jason Grout

[Harald Schilly]

On Nov 19, 6:15 pm, Jason Grout <jason-s...@creativetrax.com> wrote:
> http://alpha.sagenb.org/home/pub/3/
>

Cool ;)

Interesting, last part seems to be correlated and the overall trend is
on our side.

But back to the topic, the advantage of MMA is that it is taught at
universities (i.e. mine) ... that's one of the things we have to hook
in: make it easy to teach sage, lower entrance level, educating
teachers, ... more on sage-marketing ;)

H

[Dr. David Kirkby]

Very nice Jason!

I did not realise the data was available in that format. The third graph in
particular certainly looks interesting, and quite depressing for Wolfram Research.

I actually quite like their product myself. I have a lot of respect for Daniel
Lichtblau and some other WRI employees. Daniel clearly knows his stuff, and does
not make exaggerated claims.

But some of the claims made by others are just silly.

I do not know what idiot wrote the one about the academic publications, but I've
ridiculed them before over this, and they still make the claim.

Dave

[Nasser Abbasi]

But Jason, You did not say which newsgroup data you used, to pull the
data from and gave no reference/link.

After looking, I see you used the

sage-support

http://groups.google.com/group/sage-support/about?hl=en

But actually you should also include sage-develop and any other sage
newsgroup which would have made sage numbers higher still.

But we are comparing apples with oranges here since Mathematics
newsgroup is moderated, which means the traffic will always be less as
a moderated group is a turn off for many users who are looking for
help (students and others as well) and can't wait for 2-3 days to get
help and wait 2-3 more days for follow up.

btw, Matlab newsgroup gets more traffic than all the sage groups and
Maple and Mathematica and probably all the R groups
(gmane.comp.lang.*) combined :)

--Nasser

[William Stein]

On Thu, Nov 19, 2009 at 3:29 PM, Nasser Abbasi <n...@12000.org> wrote:
>
>
> On Nov 19, 11:15 am, Jason Grout <jason-s...@creativetrax.com> wrote:
>> Harald Schilly wrote:
>> > Hi, about the publications ... big lol!
>>
>> > better numbers:
>> >http://groups.google.com/group/sage-support/about> Archive table
>> >http://groups.google.com/group/comp.soft-sys.math.mathematica/about>
>> > Archive table
>>
>> > Homework: Create nice plots ;)
>>
>> Okay:
>>
>> http://alpha.sagenb.org/home/pub/3/
>>
>> Jason
>>
>
> But Jason, You did not say which newsgroup data you used,  to pull the
> data from and gave no reference/link.
>
> After looking, I see you used the
>
> sage-support
>
> http://groups.google.com/group/sage-support/about?hl=en
>
> But actually you should also include sage-develop and any other sage
> newsgroup which would have made sage numbers higher still.

That would be interesting. I can see why he restricted to
sage-support though, to exclude developer discussions (like this one),
which for Mathematica would be done in the hall (or mailing list) at
WRI's illustrious office building.

>
> But we are comparing apples with oranges here since Mathematics
> newsgroup is moderated, which means the traffic will always be less as
> a moderated group is a turn off for many users who are looking for
> help (students and others as well) and can't wait for 2-3 days to get
> help and wait 2-3 more days for follow up.

The Sage mailing lists are also moderated. It's just that we're way
better at it :-)

> btw, Matlab newsgroup gets more traffic than all the sage groups and
> Maple and Mathematica and probably all the R groups
> (gmane.comp.lang.*) combined :)

How many messages per month to the matlab newsgroups?

-- William

[Dr. David Kirkby]

William Stein wrote:

>>> http://alpha.sagenb.org/home/pub/3/
>>>
>>> Jason
>>>
>> But Jason, You did not say which newsgroup data you used, to pull the
>> data from and gave no reference/link.
>>
>> After looking, I see you used the
>>
>> sage-support
>>
>> http://groups.google.com/group/sage-support/about?hl=en
>>
>> But actually you should also include sage-develop and any other sage
>> newsgroup which would have made sage numbers higher still.

I think the source should be added to the graph, so someone knows roughly what
they are looking at, without having to read a long post. Perhaps a link to the
data on the Google site, would enhance it somewhat.

> That would be interesting. I can see why he restricted to
> sage-support though, to exclude developer discussions (like this one),
> which for Mathematica would be done in the hall (or mailing list) at
> WRI's illustrious office building.

To add is sage-devel would not give a level playing field.

>> But we are comparing apples with oranges here since Mathematics
>> newsgroup is moderated, which means the traffic will always be less as
>> a moderated group is a turn off for many users who are looking for
>> help (students and others as well) and can't wait for 2-3 days to get
>> help and wait 2-3 more days for follow up.
>
> The Sage mailing lists are also moderated. It's just that we're way
> better at it :-)

Even WRI employees can't post to there without the approval of the moderator!

BTW, Steve, the moderator runs sunfreeware.com.

>> btw, Matlab newsgroup gets more traffic than all the sage groups and
>> Maple and Mathematica and probably all the R groups
>> (gmane.comp.lang.*) combined :)
>
> How many messages per month to the matlab newsgroups?
>
> -- William

In my experience in engineering, MATLAB does get used a lot more than
Mathematica, so I am not at all surprised of this.

In another post, Nassar poined out this link. If you have 18 minutes to spare,
it is worth listening to.

http://blog.wolfram.com/2009/11/12/the-rd-pipeline-for-mathematica/#more-2172



Dave

[William Stein]

On Thu, Nov 19, 2009 at 4:16 PM, Dr. David Kirkby
<david....@onetel.net> wrote:
> In another post, Nassar poined out this link. If you have 18 minutes to spare,
> it is worth listening to.
>
> http://blog.wolfram.com/2009/11/12/the-rd-pipeline-for-mathematica/#more-2172
>
> Dave
>

I think *everybody* should read that, so people understand precisely
what sort of ego-maniac self-aggrandizing opponent the Sage project is
up against.

Also relevant is
http://blog.wolfram.com/2009/11/19/future-directions-of-wolframalpha/
which lays out a vision of a future in which all mathematical-related
algorithms are implemented only in a closed webserver (Wolfram Alpha).
In which one person absolutely controls how everybody interprets and
thinks about mathematical computation.

-- William

[David Kirkby]

2009/11/20 William Stein <wst...@gmail.com>:

> On Thu, Nov 19, 2009 at 4:16 PM, Dr. David Kirkby
> <david....@onetel.net> wrote:
>> In another post, Nassar poined out this link. If you have 18 minutes to spare,
>> it is worth listening to.
>>
>> http://blog.wolfram.com/2009/11/12/the-rd-pipeline-for-mathematica/#more-2172
>>
>> Dave
>>
>
> I think *everybody* should read that, so people understand precisely
> what sort of ego-maniac self-aggrandizing opponent the Sage project is
> up against.

The text is just a transcript of the video - personally I prefer to
see & hear the video. Takes less effort on my part.

Steven Wolfram will never win any awards for his modesty. He has the
biggest ego of any person I can think of. I guess he must be pretty
fit to carry around the weight of his huge ego whereever he goes!

Dave

[Nasser Abbasi]

On Nov 19, 5:34 pm, William Stein <wst...@gmail.com> wrote:


> How many messages per month to the matlab newsgroups?
>
>  -- William

About 7,000 to 8,000.

I used data from Google newsgroups to do some plots comparing
newsgroup traffic trends over the years for the above groups.

http://12000.org/my_notes/maple_mma_matlab_trends/CAS_trends.html

Hope you find it interesting.

--Nasser

[kstueve]

<Begin copy-paste>
Innovation: The dizzying ambition of Wolfram Alpha

New Scientist Tech, Nov. 17, 2009

Stephen Wolfram wants Wolfram Alpha to generate knowledge of its own.

Alpha has been exposed to more utterances than a typical child would
hear in learning a new language, allowing it to get smarter at
understanding how people phrase their requests, he says.

"You'll be able to ask it a question, and instead of it using
knowledge that came out of a method invented 50 years ago it will
invent a new method on the fly to answer it."

http://www.newscientist.com/article/dn18163-innovation-the-dizzying-ambition-of-wolfram-alpha.html
(from http://www.kurzweilai.net/news/frame.html?main=news_single.html?id%3D11413)

Bing teams up with Wolfram Alpha

http://news.bbc.co.uk/2/hi/technology/8356217.stm

</end copy-paste>

I think competition is a good thing because it makes everyone work
harder to be best. I would like to see Sage accessible through
Google. =)
Kevin Stueve

On Nov 19, 3:42 am, "Dr. David Kirkby" <david.kir...@onetel.net>
wrote:

> http://sage.math.washington.edu/home/kirkby/Mathematica-vs-Sage/mathe...

>
> The number of new posts in this period, which was about 10 days, but I can't be
> sure exactly, are:
>
> comp.soft-sys.math.mathematica 194
> sage-s...@googlegroups.com 113
>
> That's a Mathematica/Sage ratio of 1.71
>
> That's not a huge difference. If Mathematica has several million enthusiastic
> users, then Sage should have more than one million, which I very much doubt!
>
> There are clearly problems with a direct comparison like this. It's clearly not
> a rigorous statistic.
>

> * Mathematica users can email supp...@wolfram.com, but with a turnaround time of

[rjf]

On Nov 19, 5:27 pm, William Stein <wst...@gmail.com> wrote:

>
> I think *everybody* should read that, so people understand precisely
> what sort of ego-maniac self-aggrandizing opponent the Sage project is
> up against.

Mathematica is an opponent of Sage? Huh? Maybe you mean competition
for Sage?


WRI presumably has a clear notion of how many licenses for Mathematica
have been sold.
Some are for multiple users, so there is probably no way of saying how
many users, exactly,
there are. Millions? probably not. But outside of WRI, even the
number of licenses is probably not known.

As for Sage, one can tell how many downloads there have been. Again,
some are for
multiple users.

But my recollection is that there are lots more people who appear to
have downloaded Maxima than Sage.
That doesn't count secondary distributions...

The number of messages on Sage-support is not a good measure, I
think. As I've pointed out,
about 75% of them seem to involve Maxima. If you add to that the
count of messages on
the Maxima mailing list, is Maxima more popular than Sage?

I think Maxima is more widely used than Sage, especially since every
Sage installation includes Maxima but
not vice versa.

If you want to believe that Sage is, in some sense, catching up to
Mathematica, you are entitled to your opinion.

I think, however, if you actually study what is in Mathematica, its
facilities and the additional
facilities in recent versions, then you would be in a much better
position to formulate an
opinion as to relative "velocities".

>
> Also relevant ishttp://blog.wolfram.com/2009/11/19/future-directions-of-wolframalpha/

> which lays out a vision of a future in which all mathematical-related
> algorithms are implemented only in a closed webserver (Wolfram Alpha).
>   In which one person absolutely controls how everybody interprets and
> thinks about mathematical computation.

If you don't like Wolfram Alpha, you don't need to use it. It appears
to be free, though not open source.

Wolfram is not my favorite person either. So what?

RJF

[Robert Bradshaw]

On Nov 21, 2009, at 8:42 PM, rjf wrote:

> On Nov 19, 5:27 pm, William Stein <wst...@gmail.com> wrote:
>
>>
>> I think *everybody* should read that, so people understand precisely
>> what sort of ego-maniac self-aggrandizing opponent the Sage project
>> is
>> up against.
>
> Mathematica is an opponent of Sage? Huh? Maybe you mean competition
> for Sage?

Wolfram Mathematica's (R) is competition, and in that competition I
think it's fair to say Wolfram is an opponent.

I prefer to see those working on Maxima as collaborators.

> WRI presumably has a clear notion of how many licenses for Mathematica
> have been sold.
> Some are for multiple users, so there is probably no way of saying how
> many users, exactly, there are.

Nope. They're guessing, we're guessing, and though their guess is
probably more accurate than ours we can probably peg it to within an
order of magnitude or so.

> As for Sage, one can tell how many downloads there have been. Again,
> some are for
> multiple users.
>
> But my recollection is that there are lots more people who appear to
> have downloaded Maxima than Sage.
> That doesn't count secondary distributions...
>

> I think Maxima is more widely used than Sage, especially since every
> Sage installation includes Maxima but
> not vice versa.

I bet you're right. I bet the number of installs and users of Windows
calculator eclipses both of them. And that's enough for many people,
but we're targeting more specialized audiences.

- Robert

[William Stein]

On Sat, Nov 21, 2009 at 8:42 PM, rjf <fat...@gmail.com> wrote:
> As I've pointed out,
> about 75% of them seem to involve Maxima.

Not even close. However, probably 75% of the messages that you are
able to understand involve Maxima.

>  If you add to that the count of messages on
> the Maxima mailing list, is Maxima more popular than Sage?

No. The Maxima mailing list had 4366 messages so far this year:

http://maxima.sourceforge.net/maximalist.html

The Sage lists have had about 20,000 messages so far this year.

William

[rjf]

On Nov 21, 9:12 pm, William Stein <wst...@gmail.com> wrote:

>
> The Sage lists have had about 20,000 messages so far this year.

I just brought up sage-support. It says on the bottom of the screen

1 - 30 of 3492

I put "maxima" in the search box and searched that group.
I got "2,640 results for maxima ". I have no simple way to search
for "messages that I am able to understand" whether they include the
word "maxima" or not.


So if you use sage-support as a measure of what percentage of users of
Sage are really using Maxima --- (I am not necessarily endorsing this
metric, but anyway...) there you are.

If you are interested in what percentage of number theorists are using
system X,Y, or Z, I don't see any similar approach.



Now Sage-devel has 34,253 messages according to the initial screen,
just now.

A search for "maxima" yields 4,830 hits.

I think that this has less to do with popularity than with the
interests of "developers" who are presumably a much smaller, if
chattier, group. Obviously the messages are not evenly distributed
amongst all 100+ included packages, even there.

I don't know how to search in the 20,000 messages "this year".

Do you count Axiom, Reduce, and Maxima as "opponents" of Sage, too??
Or do you only count commercial programs?

RJF

[Harald Schilly]

On Nov 22, 5:08 pm, rjf <fate...@gmail.com> wrote:
> I put "maxima" in the search box and searched that group.
> I got "2,640  results for maxima ".

google groups search results aren't topics, they seem to be messages.
so, if the word maxima is in the subject line, it counts multiple
times.
i.e. "memory limit for maxima " comes up 6 times, etc.

for the sake of interest i searched for other keywords:

sage 32700
maxima 2640
mathematica 1290
matlab 564
magma 526
c++ 440
simplify 551
solve 2000
symbolic 2450
numeric 249
install 4190
windows 1690

> Do you count Axiom, Reduce, and Maxima as "opponents" of Sage, too??

I think we all see it like robert: "I prefer to see those working on
Maxima as collaborators."

H