philosophical question
Unknown
ned
Denis Gilbert
> What might MATLAB be like in 20 or 30 years?
>
> ned
>
It's hard to answer this question withouth first asking what will computers and their operating systems be like in 20 to 30 years.
I guess we can expect voice command capability among other things. Again some people will love it and others will hate it because
of the memory and CPU requirements.
Fun question...
Looking forward to read other posts, Denis.
Nathan Cahill
>
> What might MATLAB be like in 20 or 30 years?
>
> ned
Hopefully it will finally be able to do what I want it to do, not what I
tell it to do.
Nathan Cahill
Khawer Masood
Thought controlled - now that would be something!
Has anyone tried voice recongition software with matlab? I have not used
any, but would imagine that one could dictate commands (or m-files) to
matlab and then run them.
Interesting question! (-:
KM
Denis Gilbert wrote:
>
> ned wrote in article <3a2e8cc...@news.mindspring.com>...
> > What might MATLAB be like in 20 or 30 years?
> >
> > ned
> >
>
Jordan Rosenthal
If it did, any office that used this feature would be one noisy room. :)
Imagine if video games also went to voice command: "Fire, Fire, Fire,
Fire...."
Jordan
Denis Gilbert
Yep, that would be something, would it not?
In the workplace good sound insulation between individual offices would become a must. Moreover, in open areas shared by several
people, matlab would need to be able to recognize the intended user's voice to avoid performing commands issued by a close
neighbour... :-)
Cheers, Denis
us
employee)
philosophical answer: your inflammatory question cannot be vectorized.
us
In article <3a2e8cc...@news.mindspring.com>,
(ned) wrote:
> What might MATLAB be like in 20 or 30 years?
Sent via Deja.com http://www.deja.com/
Before you buy.
Johan Kullstam
> ned wrote in article <3a2e8cc...@news.mindspring.com>...
> > What might MATLAB be like in 20 or 30 years?
> >
> > ned
> >
>
> It's hard to answer this question withouth first asking what will computers and their operating systems be like in 20 to 30 years.
>
> I guess we can expect voice command capability among other things. Again some people will love it and others will hate it because
> of the memory and CPU requirements.
ack dear god no. but it has *nothing* to do with memory or CPU requirements.
1) we already have enough people here incessantly talking on the phone
or chatting with their neighbor. if everyone had to talk to the
computer there'd be no way to think at all.
2) speech is hard to make precise. i do not know how you would go
about programming by dictation. i mean just think about the pain --
A=B+3.0; ay equals bee plus three point oh semicolon next line
B=4*B; bee equals four times bee semicolon next line
[crap i meant to make that 3.0 a 5.0]
uh, go back to that line with the "ay equals" bit. one more one
more no, back, make the 3.0 be 5.0 ....
--
J o h a n K u l l s t a m
[kull...@ne.mediaone.net]
sysengr
Steve Eddins
> ned wrote:
> >
> > What might MATLAB be like in 20 or 30 years?
> >
> > ned
>
> Hopefully it will finally be able to do what I want it to do, not what I
> tell it to do.
I've been working on mindread.m. I'm not sure when it'll be ready to
ship, though.
;-)
Steve
John Williams
> >
> > ned
> >
>
>
> I guess we can expect voice command capability among other things. Again some people will love it and others will hate it because
> of the memory and CPU requirements.
Am I the only person in the world who questions the widespread
feasibility of voice operation for just one reason:
already most of us work in "cubicle farms" or open plan offices -
imagine how impossible it would be if every single person in the office
was sitting there all day, talking to their computer! It would be a
nightmare!
It's fine for a single individual in a private office, but that's not
the way things usually work these days, and that trend doesn't seem to
be changing..
Any opinions?
John
Greg Wolodkin
Denis Gilbert <Denis.Gilb...@dfo-mpo.gc.ca> wrote:
>Jordan Rosenthal <j...@ece.gatech.edu> wrote in article <90m735$ecu$1...@news-int.gatech.edu>...
>>
>>
>> Imagine if video games also went to voice command: "Fire, Fire, Fire,
>> Fire...."
>>
>> Jordan
>
>Yep, that would be something, would it not?
>
>In the workplace good sound insulation between individual offices would become a must. Moreover, in open areas shared by several
>people, matlab would need to be able to recognize the intended user's voice to avoid performing commands issued by a close
>neighbour... :-)
BANG RM STAR! BANG RM STAR!
CLEAR ALL! CLOSE ALL FORCE!
EIG OF RAND OF ONE MILLION!
Oh the pranksters will have a field day! :) :)
Cheers-
Greg
Stefan Stoll
> What might MATLAB be like in 20 or 30 years?
us wrote:
> philosophical answer: your inflammatory question cannot be vectorized.
Now this prompted me to ask Matlab
>> why can the question not be vectorized
??? Error using ==> why
Too many input arguments.
>> why
Can you rephrase that?
That's disillusioning! Obviously Matlab still has a *long* way
to go before deserving the attribute "intelligent" ;-)
So let's wait for the next release, incorporating
an interactive Matlab feature&fluff forecasting tool.
--sTefan
Peter J. Acklam
> I've been working on mindread.m. I'm not sure when it'll be
> ready to ship, though. ;-)
Would there be a companion function, mindwrite.m?
Hopefully, it would only open the mind for _appending_...
:-)
Peter
--
function f,f=fopen(which(mfilename));.......................Just
while ~feof(f),l=fgetl(f);fprintf(.......................another
'%s ',l(1+max(find(l=='.')):end)),........................Matlab
end,fclose(f);fprintf('\n')%..............................hacker
Peter J. Acklam
> What might MATLAB be like in 20 or 30 years?
Khawer Masood <kha...@wayahead.net> writes:
> Thought controlled - now that would be something!
Knowing how my mind sometimes tends to wander, without me being
aware of it until after a while, that just might not be such a
good idea... :-)
Nathan Cahill
> I'd be happy if we just had a mindopen.m and made everyone in the world
> use it!
>
Hmmmm....
>> mindopen('schwarz')
ans =
-1
??????????????????????
:)
Joe Skrap
> What might MATLAB be like in 20 or 30 years?
>
> ned
Fast? :->
/K
Sesha Sai Vaddi
30yrs is a voice recognition software......yeah but there was this
interesting comment from someone "Hopefully it will finally be able to do
what I want it to do, not what I tell it to do" i would like to add "ONLY
MUCH FASTER"........
I a graduate student working in the area of dynamics and control and my
vision for matlab in the next 20 to 30yrs(which is an astronomical figure)
wd be the following
"Given a piece of metallic junk and some sensor outputs attatched to it,
matlab shud be able to direct a complete model identification procedure,
suggest and design possible control strategies(along with the associated
estimation schemes), and be in a position to test them on the junk(whatever
it is) given the actuators"......in fact the dynamical system need not jsut
be a piece of metallic junk, cd be an abstract economics model as well, and
all we need are the sensor outputs or measurements..........
Sesha Sai Vaddi.
Hecker
That's really not something to wish for from Matlab; that's something to wish
for from people like yourself. Matlab can only implement the technology that
control theorists and practictioners create. But then, maybe I'm being too
literal w/this thread.
What matlab can do, has done, and hopefully will continue to do is improve the
methods of going from top-level, customer driven requirements to integrated,
validated, verified, and certified software, hardware, and middle-ware. To get
an idea of this, contrast simulink with writing an equivalent model in a matlab
script or in fortran. It's pretty easy to argue that the simulink GUI leads to
higher productivity (and a much more enjoyable job). Now autocode your model
(which you can) and automatically trace your code to your requirements (which
isn't really there yet) and automatically test the software to your cert
authorities standards (no way!), we could spend more time developing ever more
complex systems.
Oh, and it would be good if it loaded faster and didn't take so much memory ;).
-Chris
Doug Schwarz
Acklam) wrote:
>Steve Eddins <edd...@mathworks.com> writes:
>
>> I've been working on mindread.m. I'm not sure when it'll be
>> ready to ship, though. ;-)
>
>Would there be a companion function, mindwrite.m?
>Hopefully, it would only open the mind for _appending_...
>
>:-)
>
>Peter
I'd be happy if we just had a mindopen.m and made everyone in the world
use it!
--
Doug Schwarz
Eastman Kodak Company
douglas...@kodak.com
Denis Gilbert
>
> I've been working on mindread.m. I'm not sure when it'll be ready to
> ship, though.
>
> ;-)
>
>
Steve,
Your mindread.m would have probably helped in Palm Beach County (Florida) to detect whether all those who voted for Pat Buchanan
actually intended to vote for Al Gore... ;-) You will find an interesting story about this in Nature Science Update at this URL:
http://helix.nature.com/nsu/001207/001207-1.html
As a possible enhancement to mindread.m, have you also started working on a parallel version (multi_mindread.m) that would be able
to read several minds at once? Secret voting procedures in large assemblies would then be vastly accelerated.
Best regards, Denis.
Steve Eddins
mindread.m itself is fully vectorized, of course.
Cheers,
Steve
Ken Davis
feature... a feature I hope we see within the next 5 years, in fact. In the
future, Matlab will know how to spread out computations across multiple
processors, either on the same machine or over a network. We shall see!
- Ken
<ned> wrote in message news:3a2e8cc...@news.mindspring.com...
> What might MATLAB be like in 20 or 30 years?
>
> ned
kcmc...@my-deja.com
>
Over the next 30 years, Matlab will be become a much simpler language
than it is now.
Certain numerical methods will become universally preferred for solving
certain problems, and these will be folded more integrally into the
language. Look at the power and simplicity of the slash operator.
New programming paradigms will simplify the bookkeeping and special-case
handling that clutters up so much existing code, in much the same way
that control structures and matrix datatypes make current matlab so much
more effective than pre-77 fortran and basic. This will be most evident
in the areas of graphics and gui. Sleek future code will make today’s
window managers and button handlers read like assembly-era gobbledygook.
Standardization of data formats for sounds, still and moving images, and
operating system calls will make most existing translation and platform-
specific routines obsolete. (TMW will eventually stop supporting Windows
to concentrate on Unix and MacOS, btw).
A “compilable loop” capability will make it practical to code simple
operations using loops. Matlab overhead (data-type checking, etc) will be
performed on all parameters at the beginning, but explicit data-typing,
strict indexing rules, and restriction to basic scalar operations will
allow the loop instructions to be compiled for efficient execution. For
example, “mean” could be coded using a compilable loop. Complied loops
will reduce the heroic acrobatics often needed to achieve vectorization,
replace mex files for many simple functions that can’t quite be
vectorized, and put a full quarter of this newsgroup out of business.
The “fluff” will blow away. The grain will remain behind.
Trust me on this,
Kevin
Sesha Sai Vaddi
(TMW will eventually stop supporting Windows
> to concentrate on Unix and MacOS, btw).
co-existing with microsoft, just last week it crashed atleast ten times for
reasons i cant still comprehend and just gives me real scares considering
its time for project submissions and paper submissions......sometime bfore
one of my friends told me that matlab has stopped supporting Mac........i
guess matlab should pursue its relations with unix and mac more vigorously
to retain its high utility and reliablity and not acquire the bad habits of
microsoft........
Sesha Sai Vaddi.
<kcmc...@my-deja.com> wrote in message news:90rgsj$463$1...@nnrp1.deja.com...
Marc Rauw
> > matlab shud be able to direct a complete model identification procedure,
> > suggest and design possible control strategies(along with the associated
> > estimation schemes), and be in a position to test them on the
junk(whatever
> > it is) given the actuators"......in fact the dynamical system need not
jsut
> > be a piece of metallic junk, cd be an abstract economics model as well,
and
> > all we need are the sensor outputs or measurements..........
> >
> > Sesha Sai Vaddi.
>
it?
>
> That would be unfair to the rest of us, who might want to do non-junk
(sic.)
> computations for example...:-)
Ok, then this feature will become the Metallic-Junk Toolbox. :-P
Marc.