On 10/5/2010 8:52 AM, Invisible wrote:
>>> Note, though, that unlike Mathematica, you cannot write "2 Pi" or even
>>> "2 pi". You must write "2 * %pi". How irritating...
>>
>> Its not so bad....
>
> No, just irritating.
>
>> And .... what's with the lack of love for Lisp?
>
> People say "Lisp is a beautiful language; everything is a list". Which
> is a nice idea. But when you actually look at the language, it's kludgy
> and ad hoc, with lots of crufty backwards-compatibility junk. No thanks.
>
>> Though using * for multiplication is a bit annoying, It's not really a
>> dealbreaker for me.
>
> It just irritates me, that's all.
>
>> It still does a reasonably good job of solving and
>> expanding equations, and the wx version makes nice output. You can also
>> grab the output in TeX format, as well.
>
> I love the way that neither Mathematica nor Maxima have yet heard of
> this cool innovation called *anti-aliased graphics*. :-P Although,
> Mathematica at least manages to do AA on text, with Maxima doesn't.
>

Maxima seems to AA text just fine. If you're using the TeX fonts for 
JSMath, All of the Symbols like summation, product, integration, etc... 
are drawn using the OS font rendering engine, so, under Windows I get 
everything rendered with ClearType.

> Maxima provides two plotting commands. The plot2d() function calls
> GNUplot to do its work (and hence has all the limitations of GNUplot),
> while the wxplot2d() function plots in the window itself (and hence is
> even more limited). In either case, it's maddeningly difficult to make
> it plot what you actually want.

Plot is really basic. Draw has more capability

> Reading through the FAQ, it quickly becomes clear that Maxima isn't
> nearly as polished as Mathematica. Lots of statements like "if you want
> to solve an equation that looks like this, use foo(), but if that
> doesn't work you could try bar(), but if the equation has a form more
> like this then bar() won't work and you have to use baz() which uses a
> different algorithm internally". In Mathematica, you'd just say "solve
> this, now" and it uses 20 years of R&D and production refinement to
> compute the answer any way it can.

Now, that /is/ irritating. I have encountered some of the rough edges 
like this already

> I haven't even looked at the possibility of programming in Maxima. Hell,
> I don't know if it's even *possible*. Mathematica is basically a
> powerful pattern-matching engine, and you can program new patterns into
> it and it will transform them. I have no idea if Maxima can do that.

Maxima is very programmable, you can go as far as defining your own 
infix operators and such, and even further than that right down to Lisp

> But most of all, when I got the trial version of Mathematica what I
> quickly discovered is that most of the equations I want to solve do not
> possess a closed-form solution anyway. And in fact, as awesome as
> Mathematica is, it turns out I don't really need a CAS for anything. And
> if I do, I can usually trick Wolfram Alpha into doing what I want. (And
> since that's powered by Mathematica, it gives you the correct answer -
> assuming you can make it comprehend the question...)

What I like about Alpha is that I can actually learn something from it 
in some cases. Maxima just produces the final result, whereas Wolfram 
Alpha will sometimes have "possible" intermediate steps. I like to know 
how something was arrived at. Rather than simply accepting the answer 
spit back. But for piddling around with mathematics and not being at the 
mercy of an internet connection, Maxima works for me.

-- 
~Mike

Post a reply to this message