[Rd] drawExpression package

Sylvain Loiseau sylvain.loiseau at unicaen.fr
Sat Jan 23 22:27:14 CET 2010


> A couple of suggestions:
> - In general, I would add corresponding R input and output just beneath the "visual" display of what happens, so that readers can try it in R and see what happens.

Yes, it lacks something like that. Perhaps a first line which copy the concrete code evaluated.

> - P. 1, para. 1: ... have a common mode, one of character, logical, and numeric (to match the order of visual items).
> 
> - P. 3, para. 4: The character mode always wins. Logical always looses.
> 
> - P. 5, para. 11: A little sentence explaining nchar() would be useful. Something like: "Count the number of characters in each string:"
> 
> - P. 6, para. 14.1: again, a little explanation: "Matrices are filled by  columns by default :" ... "But you can change it :"
> 
> - P. 7, para. 14.2: "Matrices have two dimensions and you must provide extractors for each of them. You first extract the rows, then the columns."

Thanks! I have corrected all that.

> - P. 9: you use merge() that creates a list, before introducing what a list is... This is confusing.

Well, it creates a data.frame I think, and data.frame aren't handled yet (they are seen as list)... I need a graphical convention for data.frame and for factor.

I'm wondering also how to represent functions that need a *symbol* as one of the arguments, for instance 

> 
> - P. 11, para. 16.1: you use regular expressions without introducing what it is. No R beginner that has not done serious programming knows what a "Regexp" is.

Well, this is right, but there is a mix of different logics. At the beginning, it was supposed to include very few text, to stress general principles of the langage, and to be like a reference card for quickly checking the behavior of a method. For instance, which column names are preserved by cbind? The first matrix column names or the second? etc. But the document is becoming rather an introductory text.

> (P.S.: Juste après avoir écrit tout ceci, je me rend compte que tu travailles à l'Université de Caen. Je suis passé par là: j'ai mis en place les nouvelles infrastructures d'élevage d'oursins à la Station Marine de Luc-sur-mer il y a une quinzaine d'années de cela.

J'y fais de la linguistique de corpus, mais précairement !

Regards,
Sylvain

--
Sylvain Loiseau
sylvain.loiseau at unicaen.fr

Université de Caen Basse Normandie
Maison de la Recherche en Sciences Humaines
MRSH - SH018
Esplanade de la Paix
Campus 1
14032 Caen Cedex



More information about the R-devel mailing list