[texhax] Footnotes: Numbering

[Uwe Lück]

At 14:57 23.12.06, Philip G. Ratcliffe wrote:
> > Has something like
> >    \let\OldFN=\footnote
> >    \renewcommand{\footnote}[1]{\unskip\OldFN{\hspace{2pt}#1}}
> > been suggested already, where one may not have to change anything
> > in the rest of the document; plus, when I decide later I like the
> > out-of-the-box look of \footnote better after all, I can undo my
> > change by commenting out one line.  [If that doesn't work as written,
> > I assume I can debug it quickly.
>
>Nope, and this is certainly one of the better ways of solving the problem so
>far suggested, since it doesn't involve changes to the document body.

According to my own observations, it also is the most common
way for modifying macros. I just wanted to hide non-LaTeX things
like \let from LaTeX readers when I introduced that \newcommand{\FN}.
(And I was not really decided on whether to present \unskip or not.)

Which approach is better certainly depends on whether
you are just starting to write a paper or whether you basically
are done and take the time to consider layout.

> > I routinely reinvent little wheels
> > (or flat tires) because it seems faster than looking up, possibly
> > having to download, and figuring out a new package.  Or is there
> > some fatal or "good LaTeX practice" flaw with this approach?]
>
>The relative speed here clearly depends on your ability to write flawless
>macros straight off or at any rate to debug them.  Those who can usually do
>and far be it from me to give such experts advice on their approach.
>
>My objection is aimed at the suggestion that everyone should go for the
>do-it-yourself approach.

Has somebody suggested this?

>For one thing, it makes package writing a bit
>pointless and, more importantly, many (the majority) of the posters to this
>list clearly do not fall into the category of expert macro writers nor
>indeed into that of would-be expert macro writers.
>
>For many simple tasks, a simple package exists and can be found with a
>minimum of effort (CTAN FAQ, package database etc.).

In the present simple case, this "minimum of effort" is already
too much for me. It would need at least ten minutes of an hour
(I did it indeed for this thread),
while the "re-inventing" is a a matter of at most one minute.
Usually I need hours or days for finding out what packages
could serve me. Sometimes the package description
is not very clear and one must do some tests.

These things also depend of the complexity of the task, say,
in the first instance, and also whether one has knowledge
of some internals -- which may depend on the subject.
(This is directed to that concept of "[`general'] ability to
write flawless macros".)

I don't think that package writers have a right to get package
users. There still are some points for writing packages:
(i) Providing help for tasks that are complex and difficult
indeed and need very good knowledge of certain things.
(ii) Hiding things like \let from LaTeX users.

Recall how a flame war in a different thread ended:
Some readers of texhax want to get a (LaTeX) solution
that works instantly with as little learning new things
as possible, others are curious to learn some more
bits of internals or TeX. And some like to read
discussions of different approaches.

Merry Christmas,

   Uwe.