It's time for matlab to incorporate Matt Fig's EXPAND

Juliette Salexa

unread,

Feb 21, 2012, 1:01:12 PM2/21/12

to

A function that's used everyday by thousands of people shouldn't need to be copied into every folder ! Of course we can 'add to path' , but we can't always do that when working on compute servers that are controlled by other administrators who won't let us do this !

Agree or disagree ?

Stephen

unread,

Feb 21, 2012, 1:45:12 PM2/21/12

to

"Juliette Salexa" wrote in message <ji0m57$sv1$1...@newscl01ah.mathworks.com>...

> A function that's used everyday by thousands of people shouldn't need to be copied into every folder ! Of course we can 'add to path' , but we can't always do that when working on compute servers that are controlled by other administrators who won't let us do this !
>
> Agree or disagree ?

Similar to expand, I have made a formal request in the past for padarray from the image toolbox to be included in base MATLAB.

http://www.mathworks.com/help/toolbox/images/ref/padarray.html

Stephen

Matt J

unread,

Feb 21, 2012, 1:52:12 PM2/21/12

to

"Juliette Salexa" wrote in message <ji0m57$sv1$1...@newscl01ah.mathworks.com>...

> A function that's used everyday by thousands of people shouldn't need to be copied into every folder ! Of course we can 'add to path' , but we can't always do that when working on compute servers that are controlled by other administrators who won't let us do this !

===============

I'm not sure you can ask MATLAB to incorporate a file just because your administrator is uncooperative...

dpb

unread,

Feb 21, 2012, 1:55:18 PM2/21/12

to

On 2/21/2012 12:52 PM, Matt J wrote:
...

> I'm not sure you can ask MATLAB to incorporate a file just because your
> administrator is uncooperative...

_Asking_ is easy; it's the receiving that'll be the trick methinks... :)

--

Juliette Salexa

unread,

Feb 21, 2012, 3:57:10 PM2/21/12

to

> I'm not sure you can ask MATLAB to incorporate a file just because your administrator is uncooperative...

True, but it would improve matlab significantly with very very little overhead !

Matt J

unread,

Feb 21, 2012, 5:24:10 PM2/21/12

to

"Juliette Salexa" wrote in message <ji10f6$6pm$1...@newscl01ah.mathworks.com>...

> > I'm not sure you can ask MATLAB to incorporate a file just because your administrator is uncooperative...
>
> True, but it would improve matlab significantly with very very little overhead !

You can of course try to make that case to TMW in an enhancement request.

Personally, I question whether the EXPAND tool is as fundamental and beneficial as you seem to think, and whether whatever you're using it for can't be done in better ways.

Duplicating data is always something you like to avoid if you can and MATLAB has been increasingly providing ways to move away from it. For example, it's much more rare these days to use REPMAT now that we have BSXFUN.

Matt J

unread,

Feb 21, 2012, 6:43:10 PM2/21/12

to

"Matt J" wrote in message <ji15ia$o8j$1...@newscl01ah.mathworks.com>...

> "Juliette Salexa" wrote in message <ji10f6$6pm$1...@newscl01ah.mathworks.com>...
> > > I'm not sure you can ask MATLAB to incorporate a file just because your administrator is uncooperative...
> >
> > True, but it would improve matlab significantly with very very little overhead !
>
> You can of course try to make that case to TMW in an enhancement request.
>
> Personally, I question whether the EXPAND tool is as fundamental and beneficial as you seem to think, and whether whatever you're using it for can't be done in better ways.

================

However, don't get me wrong. Matt Fig was a good programmer and the EXPAND tool would have its uses when code optimization is not a priority...

Juliette Salexa

unread,

Feb 21, 2012, 8:43:10 PM2/21/12

to

"Matt J" wrote in message <ji15ia$o8j$1...@newscl01ah.mathworks.com>...

> Duplicating data is always something you like to avoid if you can and MATLAB has been increasingly providing ways to move away from it. For example, it's much more rare these days to use REPMAT now that we have BSXFUN.

BSXFUN is absolutely amazing, I won't argue with that.

Accelereyes Jacket has been able to speed up BSXFUN by a factor of over 1000 on a GPU, for more than TWO YEARS now, and Matlab hasn't caught up ... I can't wait until they do, because BSXFUN is the main part of my biggest program.

Being a big fan of BSXFUN, I still use REPMAT and EXPAND very often because BSXFUN requires a lot more thinking, while REPMAT and EXPAND are very intuitive. 99% of the time REPMAT and EXPAND are not the bottleneck of a large scale computation.

"However, don't get me wrong. Matt Fig was a good programmer and the EXPAND tool would have its uses when code optimization is not a priority..."

Why in the past tense ? Has he stopped programming ?
Actually now I notice that he stopped posting on here exactly 1 year ago ... he used to post everyday .... I love that guy.

Matt J

unread,

Feb 22, 2012, 9:22:11 AM2/22/12

to

"Juliette Salexa" wrote in message <ji1h7e$skt$1...@newscl01ah.mathworks.com>...

>
> Why in the past tense ? Has he stopped programming ?

===========

I just meant that I haven't seen him posting in the NG for a while.

Steven_Lord

unread,

Feb 22, 2012, 9:52:36 AM2/22/12

to

"Matt J " <mattja...@THISieee.spam> wrote in message
news:ji2tmj$aso$1...@newscl01ah.mathworks.com...

It looks like he and Walter have decided to switch to MATLAB Answers instead
of CSSM.

http://www.mathworks.com/matlabcentral/answers/contributors

--
Steve Lord
sl...@mathworks.com
To contact Technical Support use the Contact Us link on
http://www.mathworks.com

AccelerEyes

unread,

Feb 22, 2012, 12:48:55 PM2/22/12

to

On Feb 22, 9:52 am, "Steven_Lord" <sl...@mathworks.com> wrote:
> It looks like he and Walter have decided to switch to MATLAB Answers instead
> of CSSM.

Juliette, thanks for the props on Jacket! Two years is a long time to
wait... perhaps you should give up hope and go with Jacket? I'll give
you a good deal, email me: john.me...@accelereyes.com.

It's a shame that people are using Answers. MathWorks deletes posts
related to Jacket from Answers, and who knows what else they are
editing behind-the-scenes. Answers is definitely not a free voice
venue. In contrast, Newsreader (which is a usenet group,
http://groups.google.com/group/comp.soft-sys.matlab/topics) is a much
more open platform where people are free to discuss MATLAB without
having MathWorks make its selective "edits".

Oliver Woodford

unread,

Feb 25, 2012, 9:02:25 PM2/25/12

to

AccelerEyes wrote:
> MathWorks deletes posts related to Jacket from Answers.

Wow. I am absolutely appalled to hear that.

Bruno Luong

unread,

Feb 26, 2012, 2:30:15 AM2/26/12

to

AccelerEyes <john.me...@accelereyes.com> wrote in message <50499c81-e660-45de...@1g2000yqv.googlegroups.com>...

> It's a shame that people are using Answers. MathWorks deletes posts
> related to Jacket from Answers, and who knows what else they are
> editing behind-the-scenes. Answers is definitely not a free voice
> venue.

I never feel Answer is a transparent dialogue exchange platform. If this was true, it's a shame.

Bruno

Nasser M. Abbasi

unread,

Feb 26, 2012, 2:57:36 AM2/26/12

to

I never understood the point of 'Answers' anyway. Dividing
resources and knowledge between different places is just plain
silly. Hard to search for things now, and we have an open and
free usenet group here already. Just keep using it.

Maple did something like this also, they created their
own forum called MaplePrimes when there is a usenet for
Maple already. Never understood what is the point of it other
than controlling the questions being asked.

All closed forums that work by registration and uses these
upvoting/downvoting and collecting points for answers are
terrible idea for many reasons and I do not like them, but
they seem to be starting up everywhere these days.

--Nasser