[Rd] [R] Semantics of sequences in R

[Martin Maechler]

>>>>> "WK" == Wacek Kusnierczyk <Waclaw.Marcin.Kusnierczyk at idi.ntnu.no>
>>>>>     on Tue, 24 Feb 2009 11:31:13 +0100 writes:

    WK> Dimitris Rizopoulos wrote:
    >> in my opinion the point of the whole discussion could be summarized by
    >> the question, what is a design flaw? This is totally subjective, and
    >> it happens almost everywhere in life. Take human languages as an
    >> example and in particular, English. I do not know the history of the
    >> English language but I can guess at some point some people decided
    >> that the past tense for "give" should be "gave" and not "gived"
    >> according to the standard rule, possibly because they thought it has
    >> better acoustic.
    >> 
    >> Is this a design flaw of English? Some might argue yes, maybe they
    >> would think "gived" does not have a that bad acoustic or they could
    >> have come up with another possibility than "gave". Does this confuse
    >> new users of English? Of course it does -- I had to spent many hours
    >> learning the past tense and past particle of the irregular verbs.
    >> Should it be changed? Then almost all existing code (i.e., English
    >> texts) should be rewritten, which I think demonstrates why some people
    >> are a bit reluctant in design changes.
    >> 
    >> To close I'd like to share with you a Greek saying (maybe also a
    >> saying in other parts of the world) that goes, for every rule there is
    >> an exception. The important thing, in my opinion, is that these
    >> exceptions are documented.

    WK> all this is true;  however, programming languages are not natural
    WK> languages, there are substantial differences, and conclusions valid for
    WK> natural languages are not necessarily valid for programming languages.

You are are right, Wacek.

However, Dimitris' comparison is still a valuable one, and I
think I know that you don't quite agree :

The S language has a long history and a relatively large user
base that goes back many years (say 20). As you know, the vast
majority are not professional programmers, not even
semi-professional ones. 
Rather  applied statisticians, scientists in many fields, also
mathematicians, most very happy about how productively they can
apply the S language by using R.
The books they have written do exist however (namely mostly collections
of "R script" files), and for almost all of them it would just
lead to considerable frustration if one of 
the "exceptions in the language" was replaced by "the rule" in a
way that makes their books contain "typos".

We very occasionally do this, i.e., back-incompatible
improvements to R, inspite, but only rarely, when we are
convinced that the costs (user frustration, need to re-write
books) seem to be outweighed by the benefits. 

I think this is one of the big differences between (S and) R
and other computer languages you've mentioned.
So, indeed, Dimitris' parabola was more to the point than you
may have appreciated.

Hmm, but now let's return to something a bit more
productive, please...

Martin