[texhax] Obsolete \centerline command used in amsbook class (Uwe L?ck)

[Jim Diamond]

On Tue, Sep 25, 2012 at 10:43 (+0200), Ulrike Fischer wrote:

> Am Mon, 24 Sep 2012 18:07:36 +0100 schrieb Philip TAYLOR:

>>> Beside this amslatex doesn't "undefine" \over. The old meaning is
>>> stored in \@@over and you can easily restore it if you want. And even
>>> if it would undefine it, you could easily save the meaning yourself
>>> before loading the package. So what?

>> It is, I think, Ulrike, a clash of two different cultures and two
>> different sets of expectations.  There are people such as Michael
>> and myself, who would be perfectly happy for formats such as
>> <whatever, including empty>LaTeX to /add/ functionality to TeX,
>> so long as such packages do not remove functionality that would
>> exist were that format not loaded;  and there are others, perhaps
>> such as yourself, who are perfectly happy to accept LaTeX as a
>> /replacement/ for the (admittedly far less sophisticated) features
>> of (Plain) TeX, and who do not care one iota if a TeX primitive or
>> TeX functionality is hidden, on the perfectly reasonable grounds that
>> <whatever>LaTeX provides a far more sophisticated alternative.

>> Now people such as Michael and I are probably members of a very small
>> minority, but to dismiss their very real concerns with a casual "So
>> what ?" is tantamount to say that their views do not matter.  If that
>> is your view, then you are perfectly entitled to hold it; but by the
>> same token, Michael and I are also entitled to hold our views, and it
>> would (IMHO) be far better if each group were to acknowledge the right
>> of the other group to hold their views without seeking to suggest that
>> such views are either wrong or irrelevant.

> No the problem is not "if I care one ioto or not", or some "pure
> TeX" against "LaTeX"-thing  but that I let other people (package &
> format writer) the right to make decisions and use all possibilities
> of TeX.

> One "functionality" of TeX is that you _can_ redefine primitives.
> This functionality is built-in. And the TeXbook make say that it is
> *dangerous* to redefine primitives, but it doesn't *forbid* it. 

> Michael wants the right to use \def and \matrix even if LaTeX or
> amslatex discourages the use of this commands, but on the other side
> he denies the right of the amslatex and latex3 writers  to do what
> they think is senseful - as far as can see on the only ground that
> it makes life for himself a bit more difficult. 

> Sorry but in my long years of using LaTeX it has happened quite
> often that some change in a package or in the format - as in the
> case of the switch from latex2.9 to latex2e -  forced me to adapt
> some documents.  And then I grumble a bit and go on, sometimes I
> tried to convice a package writer that the change was to the worse
> but I didn't call such changes "inexcusable": If I couldn't convince
> the responsable I write my way around the problem.  

I find your analogy and your argument uncompelling, to say nothing of
petulant.  If a LaTeX package writer updates a package and something
changes, one might presume that he/she did so because he/she was
fixing a bug or improving the capability of the package.  And thus the
change is not only excusable but maybe even a Good Thing.

On the other hand, redefining \over or \def strikes me as gratuitously
authoritarian.  The argument that someone might use \def to clobber
some part of a LaTeX package is weak; a "pure" LaTeX user won't know
about \def, and someone who knows enough (plain) TeX to know about
\def probably realizes the danger of doing so, and can make a grown-up
decision whether the risk of using \def is worth the gain.  (Yes,
there is probably some small percentage of people who have learned
about \def and haven't figured out that using it willy-nilly is
dangerous, but that is all part of the learning experience.)

Cheers.
                              Jim