I do not like the article that much (not even in a Facebook sense). For orientation
there might be http://cs.carleton.edu/faculty/dlibenno/papers/thesis/thesis.pdf
which is Liben-Nowell's original as well as the excerpt form Kleinberg's book
http://www.cs.cornell.edu/home/kleinber/networks-book/networks-book-ch20.pdf,
one Guru, and certainly other papers - regarding social relevance and concerns
(yes, the one with the Nevanlinna price 2006)

There is an extra section Computer Science > Social and Information Networks at
arxiv.org (and in other science as well).


That is the 'framework' for data explorers like Facebook, Google, Yahoo and
more, like Amazon et al. And their 'clients'.

And one of the reasons why I de-activate trackers here at the board, like
Facebook Connect + FB Social plugins + Omniture and even more (which I
do not see any more due to activated black lists).

Fascinating anyway. But do you know Dürrenmatt's "The Physicists" and the
oldish Orwell "1984" and similar? Then you can guess my political attitude.


And I do not even want to provide subsequent and essentially related infos
how at deeper technical level these methods are used in 'non-free' countries:

The above are only used to spy out people for economic reasons, fortunately.

Whatever: the originals (given above) are more instructive, technically.

I have given up to comment on Maple's marketing.

Anyway: as soon as I use my browser with reasonable security & privacy settings
it is not even possible to view the html-versions of the contributions. So what ...

Just search for "polynomial root". Maple can do much more than what I see. And
I am sure: there are more contributions then those shown up (I see 6).

Actually I rarely look at the center: they provide *.mw and that can not be properly
converted to *.mws.

PS and just to be more specific: say you are a 'newbie' and wnat to see what you
can find about Maple and solving polynomials, you are after 'polynom root' (and
may even have a typo) but are about to buy the product, just want to know what
is in the 'public domain' beyond the product - to compare with MMA.

Entering that will return: nothing (even without typo),
http://www.maplesoft.com/applications/search.aspx?term=polynom+root

While using Google (which not even needs correct spelling) and only restricting to
the very domain gives:

http://www.google.de/search?as_q=polynom+root+maple+application+center&as_epq=&as_oq=&as_eq=&as_nlo=&as_nhi=&lr=&cr=&as_qdr=all&as_sitesearch=http%3A%2F%2Fwww.maplesoft.com%2Fapplications

(sorry for the long link)

It works for me in Maple 16, both classical and standard interface (on Win XP, SP3)

May be you upload your sheet (and ini-file, if you have created one) and also talk
about operating system and graphic device.

It works for me in Maple 16, both classical and standard interface (on Win XP, SP3)

May be you upload your sheet (and ini-file, if you have created one) and also talk
about operating system and graphic device.

There is a blank between the function and the brackets and
the standard interface uses a multiplication sign (no, that was
not intended by me ... it is an annyoing feature)

PS: may be it is better to create the Array only once and just
fill the entries (by the way: yours need ~ 580 sec on my PC)

There is a blank between the function and the brackets and
the standard interface uses a multiplication sign (no, that was
not intended by me ... it is an annyoing feature)

PS: may be it is better to create the Array only once and just
fill the entries (by the way: yours need ~ 580 sec on my PC)

Yes, that's what I meant.

The warning is correct: the userdefined function is not part of the language C

Yes, that's what I meant.

The warning is correct: the userdefined function is not part of the language C

Thx for sketching it - did not understand much (I do not know these techniques), but do not mind.

Hm ... not testing (what is Maple version + Operating System + ...).

I am running out of ideas, but if you know in advance, what has to be your input z
(need it actual be complex - that is quite expensive; what is the actual abs(z) you
need etc ) for those Bessel functions, then you could do some 'processing in bulks':

By that I mean you input arrays of your z (say in sizes of 2^10 or so) and work on
them in a *single* call, i.e. loop within the compiled program.

The idea: calling functions is very costly in Maple (lots of conversions and checks)
and you talked about 500^2 = 250000 calls.

And provide the output arrays: if properly done, they are not copied, actually only
some memory adress + datatype is transmitted.

PS on the computation of the Bessel K: this may be as in Abramowitz & Stegun by
computing the according BesselI first (approximately) and then involving log(z)
to approximately compute the difference to BesselK. That way one can improve the
values of the coeffcients used (one does not save something with short numbers).

Thx for sketching it - did not understand much (I do not know these techniques), but do not mind.

Hm ... not testing (what is Maple version + Operating System + ...).

I am running out of ideas, but if you know in advance, what has to be your input z
(need it actual be complex - that is quite expensive; what is the actual abs(z) you
need etc ) for those Bessel functions, then you could do some 'processing in bulks':

By that I mean you input arrays of your z (say in sizes of 2^10 or so) and work on
them in a *single* call, i.e. loop within the compiled program.

The idea: calling functions is very costly in Maple (lots of conversions and checks)
and you talked about 500^2 = 250000 calls.

And provide the output arrays: if properly done, they are not copied, actually only
some memory adress + datatype is transmitted.

PS on the computation of the Bessel K: this may be as in Abramowitz & Stegun by
computing the according BesselI first (approximately) and then involving log(z)
to approximately compute the difference to BesselK. That way one can improve the
values of the coeffcients used (one does not save something with short numbers).

@acer 

Thank you pointing to that subtle thing ...

At least the 'Bessel' routine uses the conventional brackets [ ] for indexing.

@acer 

Thank you pointing to that subtle thing ...

At least the 'Bessel' routine uses the conventional brackets [ ] for indexing.

The reason is not clear, and a kind of guess as well.

The discussion started at the usenet in the Maple group, "Maple slow's down the
computation when the stored amont of data grows in the memory", where I got
timing results different from Pekov

https://groups.google.com/forum/?fromgroups#!topic/comp.soft-sys.math.maple/cUz0JLoh6oA[1-25]
(sorry for that ugly interface)

The reason is not clear (Maple version, Operating System, physical memory etc)

For his overall task - which needs many evaluation, say over 10^6, I also can
imagine that the Operation System is involved, say by swapping. But I want to
stay away from those 'low level' considerations, if possible.

The reason is not clear, and a kind of guess as well.

The discussion started at the usenet in the Maple group, "Maple slow's down the
computation when the stored amont of data grows in the memory", where I got
timing results different from Pekov

https://groups.google.com/forum/?fromgroups#!topic/comp.soft-sys.math.maple/cUz0JLoh6oA[1-25]
(sorry for that ugly interface)

The reason is not clear (Maple version, Operating System, physical memory etc)

For his overall task - which needs many evaluation, say over 10^6, I also can
imagine that the Operation System is involved, say by swapping. But I want to
stay away from those 'low level' considerations, if possible.

So what do *actually* need ? May be it would help to say more
about the context instead of restricting to technical details only