|
|
| Re: Interesting article in Nature [message #83021 is a reply to message #83018] |
Thu, 31 January 2013 07:21   |
Russell Ryan
Messages: 122
Registered: May 2012
|
Senior Member |
|
|
I generally acknowledge the "tool" in professional papers if the tool is either non-standard (like I've developed some optimization routines in IDL) or am using non-standard settings of standard routines. In some cases the author(s) of the code(s) publish papers on them, so citing my source is easy. But in the case of IDL, I generally just say something like: "... for this we implement [some algorithm] in IDL\footnote{GIVE EXELIS URL}..." Though, I've never published in Nature...
On Thursday, January 31, 2013 10:15:17 AM UTC-5, Mark Piper wrote:
> On Wednesday, January 30, 2013 4:24:09 PM UTC-7, Paulo Penteado wrote:
>
>> It is not everyday that choosing IDL over other languages gets
>
>>
>
>> discussed in Nature:
>
>>
>
>>
>
>>
>
>> "The algorithms to be incorporated were varied, and included codes for
>
>>
>
>> estimating snow coverage, grain size and absorption of solar radiation
>
>>
>
>> by dust and black carbon. They had been written in IDL, a specialized
>
>>
>
>> programming language used by many researchers. Geographers, remote
>
>>
>
>> sensing experts and software programmers contributed.
>
>>
>
>> Most computer scientists would assume that such a system would take
>
>>
>
>> years, not weeks, to develop. The algorithms would presumably have to
>
>>
>
>> be rewritten in a standard language such as C++, Java or Python, or
>
>>
>
>> one that could run on a fast computer system or infrastructure, such
>
>>
>
>> as Google’s MapReduce model.
>
>>
>
>> But, in my experience, there is no need to rewrite scientific
>
>>
>
>> algorithms for bigdata systems. Rewriting only increases the barriers
>
>>
>
>> to communication between scientists and computer engineers. Rewriting
>
>>
>
>> can also introduce costly errors."
>
>>
>
>>
>
>>
>
>> From
>
>>
>
>>
>
>>
>
>> http://dx.doi.org/10.1038/493473a
>
>
>
> Thank you, Paulo, for linking to this article. I love to post articles such as this internally at VIS to try to help people understand how and why IDL is important to us who use it.
>
>
>
> mp
|
|
|
|
| Re: Interesting article in Nature [message #83022 is a reply to message #83021] |
Thu, 31 January 2013 07:15 |
Mark Piper
Messages: 198
Registered: December 2009
|
Senior Member |
|
|
On Wednesday, January 30, 2013 4:24:09 PM UTC-7, Paulo Penteado wrote:
> It is not everyday that choosing IDL over other languages gets
>
> discussed in Nature:
>
>
>
> "The algorithms to be incorporated were varied, and included codes for
>
> estimating snow coverage, grain size and absorption of solar radiation
>
> by dust and black carbon. They had been written in IDL, a specialized
>
> programming language used by many researchers. Geographers, remote
>
> sensing experts and software programmers contributed.
>
> Most computer scientists would assume that such a system would take
>
> years, not weeks, to develop. The algorithms would presumably have to
>
> be rewritten in a standard language such as C++, Java or Python, or
>
> one that could run on a fast computer system or infrastructure, such
>
> as Google’s MapReduce model.
>
> But, in my experience, there is no need to rewrite scientific
>
> algorithms for bigdata systems. Rewriting only increases the barriers
>
> to communication between scientists and computer engineers. Rewriting
>
> can also introduce costly errors."
>
>
>
> From
>
>
>
> http://dx.doi.org/10.1038/493473a
Thank you, Paulo, for linking to this article. I love to post articles such as this internally at VIS to try to help people understand how and why IDL is important to us who use it.
mp
|
|
|
|
|
|
| Re: Interesting article in Nature [message #83026 is a reply to message #83023] |
Thu, 31 January 2013 02:33 |
lecacheux.alain
Messages: 325
Registered: January 2008
|
Senior Member |
|
|
Le jeudi 31 janvier 2013 10:38:19 UTC+1, Fab a écrit :
> Hi Paulo,
>
>
>
> Thanks for the link
>
>
>
> By the way, in papers it is almost never mentioned in the
>
> acknowledgements or so with which tool the graphics have been made. Is
>
> IDL (i.e. exelis) for example officially requiring it's users to mention
>
> IDL in their manuscripts?
>
>
>
> Fab
>
>
>
> On 01/31/2013 12:24 AM, Paulo Penteado wrote:
>
>>
>
>> It is not everyday that choosing IDL over other languages gets
>
>> discussed in Nature:
>
>>
>
>> "The algorithms to be incorporated were varied, and included codes for
>
>> estimating snow coverage, grain size and absorption of solar radiation
>
>> by dust and black carbon. They had been written in IDL, a specialized
>
>> programming language used by many researchers. Geographers, remote
>
>> sensing experts and software programmers contributed.
>
>> Most computer scientists would assume that such a system would take
>
>> years, not weeks, to develop. The algorithms would presumably have to
>
>> be rewritten in a standard language such as C++, Java or Python, or
>
>> one that could run on a fast computer system or infrastructure, such
>
>> as Google’s MapReduce model.
>
>> But, in my experience, there is no need to rewrite scientific
>
>> algorithms for bigdata systems. Rewriting only increases the barriers
>
>> to communication between scientists and computer engineers. Rewriting
>
>> can also introduce costly errors."
>
>>
>
>> From
>
>>
>
>> http://dx.doi.org/10.1038/493473a
>
>>
Indeed an interesting paper.
The author, who is a scientist, express that the way in which science data are transformed by processing (by using IDL, Matlab, Python or any low level languages) becomes now less important that the way in which the resulting processed data can be archived, distributed or even merged together.
Sure that general and efficient ways for managing multi source data remain to be found (e.g. semantic web research). Nevertheless, regarding its future (and present!) scientific use, I feel important that IDL will be kept fully capable to access all sorts of existing or future shared data structures.
alain.
|
|
|
|
| Re: Interesting article in Nature [message #83027 is a reply to message #83026] |
Thu, 31 January 2013 01:38 |
Fabzi
Messages: 305
Registered: July 2010
|
Senior Member |
|
|
Hi Paulo,
Thanks for the link
By the way, in papers it is almost never mentioned in the
acknowledgements or so with which tool the graphics have been made. Is
IDL (i.e. exelis) for example officially requiring it's users to mention
IDL in their manuscripts?
Fab
On 01/31/2013 12:24 AM, Paulo Penteado wrote:
>
> It is not everyday that choosing IDL over other languages gets
> discussed in Nature:
>
> "The algorithms to be incorporated were varied, and included codes for
> estimating snow coverage, grain size and absorption of solar radiation
> by dust and black carbon. They had been written in IDL, a specialized
> programming language used by many researchers. Geographers, remote
> sensing experts and software programmers contributed.
> Most computer scientists would assume that such a system would take
> years, not weeks, to develop. The algorithms would presumably have to
> be rewritten in a standard language such as C++, Java or Python, or
> one that could run on a fast computer system or infrastructure, such
> as Google�s MapReduce model.
> But, in my experience, there is no need to rewrite scientific
> algorithms for bigdata systems. Rewriting only increases the barriers
> to communication between scientists and computer engineers. Rewriting
> can also introduce costly errors."
>
> From
>
> http://dx.doi.org/10.1038/493473a
>
|
|
|
|
| Re: Interesting article in Nature [message #83165 is a reply to message #83021] |
Fri, 08 February 2013 11:01 |
penteado
Messages: 866
Registered: February 2018
|
Senior Member
Administrator |
|
|
I have resolved to, whenever possible, publish with a paper all the
data and source code I used to obtain that paper's results.
That way anyone can download everything, replicate the results, and
see how it was done. It also causes the tools and languages used to be
mentioned somewhere. In the paper itself, I put something along the
lines of "The data and the IDL source code used to produce the results
shown here are available at xxx".
Just as arXiv requests the LaTeX source files and there are archives
to publish the data produced for the paper, I think publishing source
code should be standard practice.
On Jan 31, 1:21 pm, rr...@stsci.edu wrote:
> I generally acknowledge the "tool" in professional papers if the tool is either non-standard (like I've developed some optimization routines in IDL) or am using non-standard settings of standard routines. In some cases the author(s) of the code(s) publish papers on them, so citing my source is easy. But in the case of IDL, I generally just say something like: "... for this we implement [some algorithm] in IDL\footnote{GIVE EXELIS URL}..." Though, I've never published in Nature...
>
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
