[R] The Perils of PowerPoint

[(Ted Harding)]

On 06-Sep-05 Mulholland, Tom wrote:
> For some reason (probably that our organisation has blocked the site) I
> could not see the original articles that prompted the post. I however
> immediately assumed that this was precipitated by Tufte and his
> comments about PowerPoint (I recall seeing a good example of PowerPoint
> on his site) http://www.edwardtufte.com/tufte/powerpoint
> 
> When this first came up I recall some dispute about the comments
> www.sociablemedia.com/articles_dispute.htm and that John Fox did
> something http://ils.unc.edu/~jfox/powerpoint/introduction.html that I
> enjoyed reading.
> 
> Other links that are lying on my computer are
> "In defense of PowerPoint"
> http://www.jnd.org/dn.mss/in_defense_of_powerp.html
> and "Does PowerPoint make you stupid?" at
http://www.presentations.com/presentations/delivery/
article_display.jsp?vnu_content_id=1000482464
>  
> Tom

Thanks, Tom, for these pointers to interesting discussions!
One must of course agree with the general comments to the effect
that the quality and merits of a presentation are the result of
choices made by the person who designed it, and not primarily due
to the software itself. It is also true that software such as
PowerPoint provides ready-made mechanisms for linking-in a great
variety of content, thereby making it -- in principle -- easier
for the designer to choose judiciously what would be best for the
result they wish to achieve and -- in principle -- to design an
outstanding presentation.

It is nevertheless still true that in practice the result is often
dreadful, for reasons which largely reside in the software (but
which take effect by virtue of user deficiency).

I tend to put this down to the provision of so-called "Wizards"
-- in reality electronic snake-oil merchants -- the protoype of
which is the dancing paper-clip masquerading as an "Office
Assistant". There are other "resources" which can have similar
effects -- "spell-checkers", "grammar-checkers", auto-formatters
which brush you aside and re-arrange your intentions and which
can be difficult to evade: indeed, one can form the impression
that it has been deliberately made difficult for users to ignore
these things and make their own choices.

In case you may wonder how I hope to bring this On-Topic, it is
as follows. The result of such things is that users' thought
and practice become software-led and software-driven. The software
is both carrot and stick. The user is the donkey.

In contrast, as software and in its implementation as a compendium
of resources and documentation, R expects users to know what they
are doing and to understand the rationale of the methods. R also
requires users to have the capability to locate necessary inforamtion
in the documentation. Indeed, one might even describe R documentation
as notoriously unintrusive!

So using R should educate users in thoughtful and judicious use of
statistical software. The same cannot be said so wholeheartedly of
S-Plus. While the latter is basically routine-equivalent to R, and
the help and menu systems properly used can also encourage judicious
use, there is nevertheless a superficial aspect which can seduce users
into a "check-box" mentality; and the printed manuals strike me as
both unclear and unduly prescriptive.

In other words, while S-Plus may tend to attract users who do not
know what to do and who expect the softare to tell them what to do
(and subsequently will not know what they have done), R will not.

This spartan environment is lean and healthy, so successful R users
will become lean and healthy! Not donkeys, but mountain-goats.
R-help is there for those who need it, and very few responses to
queries have been at all superficial. Often it is clear that
respondents themselves have had to think before being able to come
up with an answer, and very often the response urges the questioner
to think! Indeed, evidence of thought on the part of the questioner
is something of a pre-requisite for getting a response.

The underlying thought behind all this is that there is something
of an under-current of disquiet in the statistical community about
"software-driven analysis", an increasingly prevalent abuse of our
subject. Occasionally it comes to the surface. Crass abuses such
as are encouraged by PowerPoint snake-oil and the like are obvious;
but once we perceive them we can be sensitised to similar but more
subtle dangers in other software. Conscious remedial effort would
be a good thing, and R seems to be an excellent vehicle for it.

Thanks for reading so far!

Best wishes to all,
Ted.


--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at nessie.mcc.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 06-Sep-05                                       Time: 14:29:26
------------------------------ XFMail ------------------------------