[ESS] Integration of pandoc using polymode+knitr

Thu Mar 27 21:58:27 CET 2014

Le 03/27/2014 04:33 PM, Yihui Xie a écrit :
> Right -- rmarkdown is not a CRAN package _yet_, but no, it is not tied
> to the RStudio company in the sense of a software package. As you can
> see, it is open source (GPL-3): https://github.com/rstudio/rmarkdown I
> do not think there is anything wrong for a company to write open
> source software packages just because it is a company instead of a
> university or a non-profit organization. There are great open source
> products from both *.com and *.org parties. I'm not sure why you are
> so concerned about the "label", or what your concerns really are, and
> we (==RStudio) will truly appreciate it if you tell us what we did
> wrong. Open source should not stop people from being rich, should it?
> (although I'm not rich...)
>
> The rmarkdown and related products are still under development.
> Eventually rmarkdown will be a CRAN package. I see your point about
> the non-R portion of source code (it is big... or is 8.5Mb really
> big?), but I guess that misses the goal. The goal is not to write more
> R (and only R) code in a package, but to have better-looking output
> with less effort in the low-level tuning of Pandoc parameters. There
> are plenty of R packages on CRAN that do not primarily contain R code,
> right? (e.g. packages for datasets, packages using C++, and why should
> packages containing more JavaScript code be second-class citizens? :)
>
> BTW, I'm happy to continue the discussion in private if this is
> considered off-topic.

Yihui and Martin: I'm no maintainer of the ESS help list, but please don't 
switch it to a private discussion. Although I don't have much to add, or 
simply comment, I think the issues discussed are also interesting for the 
general ESS audience, from a practical (knitr vs. rmarkdown), technical 
(size of rmarkdown and inclusion of Javascript), and philosophical (company 
vs. non-profit) perspectives. Both your points seems perfectly valid to me 
at this point, and they help choosing a tool using the best information 
available.

At least I would be interested.

Thanks,
Mathieu.

> Regards,
> Yihui
> --
> Yihui Xie <xieyihui at gmail.com>
> Web: http://yihui.name
>
>
> On Thu, Mar 27, 2014 at 12:54 PM, Martin Maechler
> <maechler at stat.math.ethz.ch> wrote:
>>>>>>> Yihui Xie <xieyihui at gmail.com>
>>>>>>>      on Thu, 27 Mar 2014 01:03:15 -0500 writes:
>>
>>      > Hi Mathieu,
>>      > Sorry for the late reply. The rmarkdown package has made a lot of
>>      > efforts in generating attractive output by default, including HTML and
>>      > PDF output, etc. For example, HTML output uses Bootstrap 3 and you can
>>      > easily switch themes.
>>
>> Well, rmarkdown is not a CRAN package... and it seems is really
>> tied to the Rstudio company.
>> It looks to me one way to have people flock to using RStudio everywhere..
>>
>>      > I know it may be a very bad idea to suggest the RStudio IDE in the ESS
>>      > mailing list,
>>
>> yes...  ;-)
>>
>>      > but I still recommend you to try it out. It is super
>>      > easy to go from an R Markdown document to other output formats. You do
>>      > not have to remember the Pandoc arguments -- there is a configuration
>>      > wizard and a little document setting button. Well, I know GUI's
>>      > probably do not make sense here...
>>
>>      > Regards,
>>      > Yihui
>>      > --
>>      > Yihui Xie <xieyihui at gmail.com>
>>      > Web: http://yihui.name
>>
>>
>>      > On Tue, Mar 11, 2014 at 9:19 PM, Mathieu Basille
>>      > <basille.web at ase-research.org> wrote:
>>      >> Dear Yihui,
>>      >>
>>      >> Le 03/11/2014 10:04 PM, Yihui Xie a écrit :
>>      >>
>>      >>> I'm glad that your problem was solved. Just as Ahmadou
>>      >>> mentioned, we
>>      >>> are heading towards rmarkdown (http://rmarkdown.rstudio.com),
>> 'we' ==?== Rstudio employees, or ??
>> I'm really getting a bit concerned here.
>> I liked Rstudio when it was *.org, and have been getting
>> increasingly nervous, now that it has been *.com ...
>>
>> As mentioned above, why should we suddenly have to work with a non-CRAN package?
>> Also, is 'rmarkdown' (as a standard) something to be defined outside of the
>> Rstudio web site any time soon?
>> I've started using *.Rmd (and Rstudio to some extent) in
>> teaching, but I'm really having concerns more and more.
>>
>>      >>>  which is a nice alternative to knitr::pandoc().
>>
>> Then, knitr is a CRAN package with GPL licence,
>> so we can rely that it will continue to be available into the future.
>> (Ok, I see that rmarkdown also has a GPL licence.. and it is on
>>   github, just not on CRAN, and it contains 8.5 MB of source code,
>>   most of which is java script ... ... )
>>
>>      >> would be fairly easy to use any other function, such as rmarkdown::render
>>      >> instead of knitr::pandoc. As for me, as I already use knitr (obviously) to
>>      >> knit, and pandoc is also installed on my system, it makes more sense to use
>>      >> what's available!
>>
>> and what is simple and modular..  The rmarkdown with all its new
>> bells and whistle and huge amounts of non-R source code, has its
>> advantages but also its drawbacks in my view.
>>
>> Martin
>>
>>      >> Or maybe I'm missing the obvious: is there a strong
>>      >> argument in favor of rmarkdown::render instead of knitr::pandoc?
>>      >>
>>      >> Mathieu.
>>

-- 

~$ whoami
Mathieu Basille, PhD

~$ locate --details
University of Florida \\
Fort Lauderdale Research and Education Center
(+1) 954-577-6314
http://ase-research.org/basille

~$ fortune
« Le tout est de tout dire, et je manque de mots
Et je manque de temps, et je manque d'audace. »
  -- Paul Éluard