I wish to discuss some points:
1. Menus and Dialog boxes
2. Discussion on the R-mailing list: R-core team contacted
3. Private Mailing List
4. Contact Persons
Erich Neuwirth wrote:
I think the idea of a flexible framework for R wizards is
a very good one.
...
In its minimal form, the framework would be based on an R function
taking some inputs and producing some results.
From the signature of the function (number and types of parameters)
and from the signature of the result description of the compound
structure representing the result) one should construct dialog boxes
querying the user for the inputs, and rules should be specified
how the components of composite result types should be
put into spreadsheet ranges.
If such a framework is available, it should be easy to add a new
function to the interface.
1. That is my idea. There need to be some dialog boxes / wizards that
would allow even a newbie perform some crucial statistical tests (like
non-parametric tests). I will deal in a later e-mail which tests I
personally think would be important to be available.
While the method in RExcel is powerful, it has some important drawbacks
in its pure RExcel form:
- a newbie does NOT know the R-syntax / commands yet
- because a meaningful error feedback is missing, newbies will have it
double hard to write correct R-commands (and debug the errors)
[However, I apologize, if there are inaccuracies, I had only limited
time to play with RExcel around.]
Although the mechanism should allow future development, I am strongly
supportive of the idea to have functioning menus / dialog boxes during
this summer.
2. That gets us to the second point:
How to embed R into Calc?
Concepts like C++, java and python were mentioned previously. Indeed,
all have strengths and drawbacks.
One important fact to note is that there are already some packages for
R, like rcom and R(D)COM which allow a foreign application to connect to
R. So we need NOT to invent everything. [see also the statistical wiki
page for further examples]
IF there is the need to ask for further details on the R-core code, I
have already opened this discussion on the R-mailing lists. Some of the
discussion took place on the R-devel mailing list (see
https://stat.ethz.ch/mailman/listinfo, or specifically the R-devel
mailing list archive, https://stat.ethz.ch/pipermail/r-devel/). The
first message in the thread can be accessed directly through
https://stat.ethz.ch/pipermail/r-devel/2007-March/045079.html .
3. Unfortunately, most of the important discussions took place on the
*R-core member list* which is NOT open to the general public. At the
same time, some of the e-mails were sent *privately*, so are NOT present
on either of the two mailing lists. I will therefore have to compile the
basic points and post them here in a later e-mail.
4. Important Persons
Maybe a last note on important contact persons, IF we need some help on
R code: Prof Brian Ripley is one of the core-team developers; Duncan
Temple Lang is the expert on Unix embedding, and both Simon Urbanek and
Duncan Temple Lang are Java gurus. They have all kindly shown interest
in this project, so I hope for a great success.
As previously stated, java is probably NOT a bad starting point and we
could later recode important things in C++. But this should be more
thoroughly discussed. I will try to emphasize some of this issues in a
later e-mail (BUT I never coded in java, so my views should be taken
with some care).
Sincerely,
Leonard
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
