Re: Summer of Code: R and Calc

Hi,

I would like to raise some important questions. Therefore, I
invite the 
whole OOo team to join this discussion. It would be wise to
have some 
good ideas concerning these issues before the actual coding
starts.


Wojciech Gryc wrote:
> the usefulness of any integration (for me) is
> based on two types of users: "advanced" and
"basic".

Erich Neuwirth wrote:
> You really want a "closed" interface for
naive users of statistics,
> probably through a menu driven interface similar to the
analysis toolkit
> in Excel.


There is indeed the need to clarify which users we want to
support 
first. I would like to have both support for basic users, as
well as 
methods for a more complete integration of R. Obviously,
this latter 
approach will be more difficult to code.

Therefore, before the real coding starts, here are my
questions:

1. What should we implement first?
A basic integration of R with some hard-coded functions
accessible from 
the Calc menu is more easily to implement and would benefit
primarily 
newbies (which are less likely able to analyse the data
using R and 
therefore would benefit the most).

A more advanced integration would benefit more experienced
users, too, 
but it is more difficult to implement.

Before answering this question, maybe the following one
should be 
discussed, too.

2. What is doable during this summer?
Is it doable to implement the more advanced version during
this summer? 
Or alternatively, what resources are needed to implement it
during this 
summer (during the Google SoC). Of course it is doable if
there are 
enough resources.

3. This implies another question:
Are there any changes necessary inside Calc and R to
implement it? 
Please bear in mind that the implementation should be 
platform-independent (aka should work on most platforms),
therefore some 
of the tasks won't be trivial.

If a lot of changes are necessary, then it will be difficult
to make it 
during this summer, without substantial coding help from
both the OOo 
team and the R core developers team. This should be taken
account of.

And now, lets open the discussion. I welcome any informed
arguing.

Sincerely,

Leonard

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribesc.openoffice.org
For additional commands, e-mail: dev-helpsc.openoffice.org

On Monday 02 of April 2007 20:47:23 Leonard Mada wrote:
> There is indeed the need to clarify which users we want
to support
> first. I would like to have both support for basic
users, as well as
> methods for a more complete integration of R.
Obviously, this latter
> approach will be more difficult to code.
I absolutely agree with you. Wizards for stastistic newbies
and functions for 
advanced users. What I believe need emphasising is not the
problem of coding 
but constructing tools for newbies. As we all know
statistics is a powerful 
decision support tool - still inferring meaningful
information from statistic 
analysis needs some information about the methods  (and
concepts - for 
example statistical significance). 

> 2. What is doable during this summer?
> Is it doable to implement the more advanced version
during this summer?
> Or alternatively, what resources are needed to
implement it during this
> summer (during the Google SoC). Of course it is doable
if there are
> enough resources.
I believe we should thing about flexible framework for
defining 
Rwizardz/Rfunctions. Having this new ones could be added
later.

> 3. This implies another question:
> Are there any changes necessary inside Calc and R to
implement it?
> Please bear in mind that the implementation should be
> platform-independent (aka should work on most
platforms), therefore some
> of the tasks won't be trivial.
As I mentioned before Java and something like 
http://stats.m
ath.uni-augsburg.de/JRI/ should do the trick 
(platform-independent).

Regards,
-- 
Krzysztof "Filo" Gorgolewski

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribesc.openoffice.org
For additional commands, e-mail: dev-helpsc.openoffice.org