Analysis of newsgroups traffic for
comp.soft-sys.math.mathematica,
comp.soft-sys.math.maple,
comp.soft-sys.matlab,
sage-devel,
sage-support,
comp.lang.fortran
comp.soft-sys.math.scilab
http://12000.org/my_notes/maple_mma_matlab_trends/index.htm
--Nasser
These are interesting graphs. Except for Matlab it seems that
everything else is trending in the noise. Of course, MMA and
Maple customers would likely have support contracts that allow
direct emails rather than newsgroups.
It would be interesting to be able to filter out every email
address that ends in ".edu" as this might give an indication
of which systems were used beyond the university. I suspect
that the numbers would be very low for all of the systems.
Matlab dominates the courses I have taken online so I also
suspect that a lot of the Matlab traffic is student driven.
These are conjectures based on my experience.
The only industrial use I have ever personally seen was to
develop Bode plots for industrial robot control systems.
What are these systems used for beyond teaching?
Tim Daly
Totally useless for anything, however...
How many of the other vendors of proprietary products use usenet as does
TMW?
Fortran, as a compile language has at multiple vendors as well as
open-source compilers. To have any hope whatever of such traffic being
of any relevance whatever one would have to include the vendor support
groups as well as the open source mailing lists at a bare minimum.
--
Interesting info. Perhaps you did this (it's hard to tell how you
generated the data entry) but you can automate the procedure of
reading the data into Mathematica considerably. For example, your
exact Mathematica data could be generated like so:
in = Import[
"http://groups.google.com/group/comp.soft-sys.math.mathematica/about",
"Data"];
data = ToExpression[Rest[Cases[in,
{"Archive", _}, Infinity][[1, 2]]]];
data[[1]] = Prepend[PadLeft[Rest[data[[1]]], 12],
First[data[[1]]]];
data[[-1]] = PadRight[Drop[data[[-1]], -1], 13];
data
Mark McClure
Not all is done in usenet, at least for Maple there is
http://www.mapleprimes.com/recent/all (unfortunately it
is in a very fashionable style & functionality ...)
wow, I had no idea I could do that in Mathematica !
I actually copied the data myself, copy/paste, one line
at a time, clean up, etc... took sometime :(
With your method, I can now add more newsgroups, and make more
graphs. The reason I did not before, becuase it was taking much
time to copy the data manually.
Thanks
--Nasser
http://www.mathworks.com/matlabcentral/fileexchange/18853-rcssm-a-pedestrian-newsgroup-stats-reader
> It would be interesting to be able to filter out every email
> address that ends in ".edu" as this might give an indication
> of which systems were used beyond the university. I suspect
> that the numbers would be very low for all of the systems.
In comp.soft-sys.matlab filtering out .edu would leave a lot of
messages. For one thing, a lot of students who post do so from gmail or
other free systems, not often from .edu . That's partly because the
matlab group gets quite a few students posting from outside North
America, but also because few of the students posting blatant homework
assignments do so from their student accounts.
Even if you eliminated postings from students, the matlab group would
still have more postings than most of the other groups. None of the
"regulars" (the volunteers who often answer questions) post from
educational institutions. You'd have to filter out *replies* to student
postings in order to start to get useful numbers.
> With your method, I can now add more newsgroups, and make more
> graphs. The reason I did not before, becuase it was taking much
> time to copy the data manually.
>
> Thanks
No problem - it's fun!
The Maple Primes group is interesting. There's an RSS feed, which is
essentially XML. Mathematica has very good XML tools so it might be
possible to access that as well. I don't know so much about RSS
queries, though.
Here's data from the Maxima discussion list:
data = Table[yeardata =
Import["http://www.math.utexas.edu/pipermail/maxima/" <>
ToString[year] <> ".txt.gz"];
yeardata =
StringCases[yeardata,
StartOfLine ~~ "From: " ~~
Shortest[x__] ~~ StartOfLine ~~
"Date: " ~~ Shortest[y__] ~~ EndOfLine];
yeardata = Table[Last[StringSplit[s, "\n"]], {s, yeardata}];
Prepend[Table[Total[StringCount[yeardata, month]], {month,
{"Jan", "Feb", "Mar", "Apr", "May", "Jun",
"Jul", "Aug", "Sep", "Oct", "Nov", "Dec"}}], year],
{year, 2006, 2010}]
Mark
fyi,
Some asked for more groups. So I added Java, Lisp, python, and updated
the charts to be easier to see. Now that the download of newsgroup
traffic is automated, it is easy for me to do this.
--Nasser
Cool. Add ruby to the mix.
:o)
That's pretty cool. So what percentage of the math and physics ones do you think are doit4mes?