r3wp [groups: 83 posts: 189283]
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

World: r3wp

[!REBOL3-OLD1]

Geomol
11-Apr-2006
[386x2]
And going further, if we take the vector road, what about matrices?
Support for multi-dimensional vectors and matrices in rebcode might 
lead to something really interesting. ;-)
Graham
11-Apr-2006
[388x2]
I think it should be handled natively
but it could get ugly quickly.
Geomol
11-Apr-2006
[390x3]
The alternative, doing vector- and matrix-math just using decimals, 
can easily be ever uglier.
Imagine having a matrix! datatype:
m: make matrix! 4x4
Then you could find the determinant like for example:
m/determinant
That would be a miracle! :-)
Suddently complicated math can be made very simple.
Henrik
11-Apr-2006
[393]
it would also be a miracle to be able to have a complex! datatype. 
I think electronic engineers would appreciate that
Graham
11-Apr-2006
[394]
How to submit a suggestion to Carl?
Pekr
11-Apr-2006
[395]
feedback or comment on blog?
Henrik
11-Apr-2006
[396x3]
or rambo?
geomol, I once had the idea that you could define ISO units as kind 
of datatypes and do scientific calculations on them
my calculator can do that, so why not have it in rebol?
Geomol
11-Apr-2006
[399]
Henrik, give an example!
Graham
11-Apr-2006
[400]
Geomol, Rambo the suggestion for a vector datatype.
Rebolek
11-Apr-2006
[401]
and what about interval computations? :)
Graham
11-Apr-2006
[402]
Let's stick with 3 dimensions.
Henrik
11-Apr-2006
[403x2]
well, I'm really not sure about the notation as I haven't given it 
too much thought, so I can't give a valid example, but if you could 
say:

>> 6 / 2
== 3

>> 6 [m] / 2 [cm]
== 300 [cm]
this is no good, but some method to attach units to numbers and variables 
and make REBOL aware of them during calculations
Graham
11-Apr-2006
[405]
>> $5.00 / 20
== $0.25
Geomol
11-Apr-2006
[406]
Henrik, ah ok. Yes, good idea!
Graham
11-Apr-2006
[407]
so we can do mixed imperial and si calculations ??
Geomol
11-Apr-2006
[408]
Something like:
ISOm6 / ISOcm2
Henrik
11-Apr-2006
[409x3]
sure, my HP48 can do that easily
as long there is some definable way to convert between them or make 
some sense from it
the acceleration of gravity would be 9.82 [m/s/s] so units themselves 
are math expressions
PeterWood
11-Apr-2006
[412]
It that something for a dialect rather then to be inlcuded in the 
core langauge?
Henrik
11-Apr-2006
[413]
peterwood, quite possibly
Graham
11-Apr-2006
[414x2]
we already have money datatypes
why not ones for mensuration
Henrik
11-Apr-2006
[416]
I think it's worth making a concept dialect for scientific math with 
units
Geomol
11-Apr-2006
[417]
Hmm, funny idea. To add 6 meters and 2 UK foot:
SIm6 + UKfoot2
Looks weird. :-)
Henrik
11-Apr-2006
[418]
actually, it would be good to make a ctx-unit to handle units exactly 
like the hp48 does
Graham
11-Apr-2006
[419]
I often have to calculate BMIs .. and most people know their height 
in feet and inches, but their weight in kgs :(
Henrik
11-Apr-2006
[420x3]
the hp48 uses an underscore to attach the number to a unit: 6_m / 
23_cm
http://www.apple.com/downloads/macosx/math_science/eurekalc.html
<-- I think that could be done with a dialect...
so many things to do!
Graham
11-Apr-2006
[423]
that calculator needs a spell checker as well!
Henrik
11-Apr-2006
[424]
well, kintetic energy is that new thing, you know...
Geomol
11-Apr-2006
[425]
Graham, I've made a RAMBO proposal for vector and matrix.
Graham
11-Apr-2006
[426x2]
how do you find the 4th part of a 4D vector?
ie. what comes after /z ?
Geomol
11-Apr-2006
[428x2]
4? :)
x, y and z can be possible, like we have first, second ...
To reach the next dimensions, integers are needed.
v: 1.0x2.0x3.0x4.0x5.0
v/x
1.0
v/5
5.0
PeterWood
11-Apr-2006
[430]
A quick google took me to the Python page - "The 4th component can 
be accessed either by the name w or t." 

http://cgkit.sourceforge.net/doc2/vec4.html
Geomol
11-Apr-2006
[431]
First, second, ... go up to tenth. We don't have eleventh. You need 
to do serie/11 to get the 11'th element.
PeterWood
11-Apr-2006
[432]
It would seem w is the popular choice - www.mathpages.com/home/kmath482.htm
Geomol
11-Apr-2006
[433]
Peter, ok idea, and I'm all for that. There just shouldn't be a limit, 
and we have to use integers to reach later parts of a vector ... 
like with series.
Graham
11-Apr-2006
[434x2]
Good to know that is solved.
don't know why we just don't start with "a" and work our way thru 
to "z"