R devel - May 2017 - SUGGESTION: R Base Packages

Dear R Developers,

I am writing for doing the following suggestions:


*         Bugs:

o   Add in Bugzilla an Area in order to request membership for bug reporting,
which shall itself include another area to introduce a little dissertation for
justifying "Why someone would like to be a member of Bugzilla" for
being able to report proper bugs, and avoid spam.

o   Now, new R serious members cannot report properly bugs.



*         That, in the R Base distribution, the following packages are included
as default (pre-installed):

o   ggplot2.

o   dplyr.

o   tidyr.



As in 2017, these are essential, in my opinion, in R programming.


*         And that, as regards R Project:

o   Improve the CRAN & R Project web pages, giving them a more modern look
and feel (Such as the Python.org webpage), in order to "sell" better R
to the everybody.

o   Improve the basic GUI distributed with R, with some of the core features of
RStudio Desktop (GPL). Or even make RStudio Desktop to replace the R default
GUI, with another name, in its executable version.

o   Enable a Suggestion web page / forum in the R Project webpage in order to
boost core R Open Innovation.

Thank you all.



	[[alternative HTML version deleted]]

Hello!

Thanks to Juan for his suggestions. I would like to voice my opinion
for and against some of these.
> o   Add in Bugzilla an Area in order to request membership for bug
reporting, which shall itself include another area to introduce a little
dissertation for justifying "Why someone would like to be a member of
Bugzilla" for being able to report proper bugs, and avoid spam.
> 
> o   Now, new R serious members cannot report properly bugs.
I too was annoyed by the state of the bug reporting system when I
first started using R, and that was even before automatic account
creation was disabled. I think the situation could be improved. I
don't think it's a huge burden to tell bug reporters to ask for an
account via the R-devel mailing list. But from what I can see the bug
tracker doesn't make it clear that this has to be done. We should at
least post a message on the main page explaining how people can create
an account. This would be less work than what Juan is suggesting, but
I think sufficient for our needs.
> *         That, in the R Base distribution, the following packages are
included as default (pre-installed):
> 
> o   ggplot2.
> 
> o   dplyr.
> 
> o   tidyr.
I think this is a terrible idea. I've only been using R for a few
years and while I find these "tidyverse" packages interesting and use
them on occasion, I've also concluded that they change too quickly to
be used in code that I want to stay working for a long time. They are
largely based on experimental concepts. Being able to change and
evolve is part of the strength of these packages. So I think putting
them in the R Base distribution would be bad for all parties.
> *         And that, as regards R Project:
> 
> o   Improve the CRAN & R Project web pages, giving them a more modern
look and feel (Such as the Python.org webpage), in order to "sell"
better R to the everybody.
> 
> o   Improve the basic GUI distributed with R, with some of the core
features of RStudio Desktop (GPL). Or even make RStudio Desktop to replace the R
default GUI, with another name, in its executable version.
> 
> o   Enable a Suggestion web page / forum in the R Project webpage in order
to boost core R Open Innovation.
I think the R project web page looks great. It's simple and it loads
quickly and doesn't try to mesmerize people. I like R's command line
interface. It completes on symbols and files and I can easily use it
with my favorite editor and run it in the terminal of my choice. I
don't think that more effort should be put into developing bloated
GUIs which try to enforce a standard way of interacting with R.

I think a "forum" or bulletin board system would be a detraction from
the project and a distraction for the project leaders. Users have
Stack Exchange - it's better than any forum we could create, and it
takes care of itself.

That's my two cents, more or less.

Thanks,

Frederick

Thank you Frederick for your comments:

They are really well justified.
> I think a "forum" or bulletin board system would be a detraction
from the project and a distraction for the project leaders. Users have Stack
Exchange - it's better than any forum we could create, and it
> takes care of itself.
An excellent idea would be to add in the R Project Webpage a link to RSeek, in
order to put all together. This can be done in a fast and easy way.
> I too was annoyed by the state of the bug reporting system when I first
started using R, and that was even before automatic account creation was
disabled. I think the situation could be improved.
> I don't think it's a huge burden to tell bug reporters to ask for
an account via the R-devel mailing list. But from what I can see the bug tracker
doesn't make it clear that this has to be done. We should at
> least post a message on the main page explaining how people can create an
account. This would be less work than what Juan is suggesting, but I think
sufficient for our needs.
Yes, clarifying in the RProject webpage that accounts could be requested for bug
reporting in Bugzilla would be a great news, but it shall be really good
justified.
> I think this is a terrible idea. I've only been using R for a few years
and while I find these "tidyverse" packages interesting and use them
on occasion, I've also concluded that they change too quickly to
> be used in code that I want to stay working for a long time. They are
largely based on experimental concepts. Being able to change and evolve is part
of the strength of these packages. So I think putting
> them in the R Base distribution would be bad for all parties.
Totally agree.
> I think the R project web page looks great. It's simple and it loads
quickly and doesn't try to mesmerize people. I like R's command line
interface. It completes on symbols and files and I can easily use it
> with my favorite editor and run it in the terminal of my choice. I
don't think that more effort should be put into developing bloated GUIs
which try to enforce a standard way of interacting with R.
I agree, but, in my opinion,  the RGUI icons  shall have a little bit more
modern look. Just a bit.

Thank you,
Juan