[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: makeinfo function execute feature

From: Robert J. Chassell
Subject: Re: makeinfo function execute feature
Date: Wed, 21 Jan 2004 14:11:53 +0000 (UTC)

    [eval Emacs Lisp in Emacs Info] is too useful a feature to abandon ....

How is it useful?  You need to explain the advantages of an act that
contradicts the fundamental form of Texinfo.  If it is a good idea, I
will support it.  But for more than 15 years, my thinking has been in
favor of modular separation of capabilities.

Fundamentally, Texinfo is a documentation format that enables you to
write one source file for various outputs, including outputs, like
books, that cannot by their very nature evaluate Emacs Lisp.

Why is it better to provide Emacs Lisp code that can be executed as
Emacs Info files rather than as .el files?  Or provided in some other
way?  That is the question you must answer.

I once thought that it would be a good idea to be able to evaluate
Emacs Lisp code in Emacs Info.  After all, this would be an easy way
to convey code to strangers.  However I was persuaded of two things:

  * It is useful for Texinfo to be a format that takes single source
    to make multiple types of output, and that be the core job that 
    Texinfo does.  This simplifies the cognitive load on humans.

  * It would be dangerous to place executable code in a document that
    cannot contain executable code in the majority of its formats.

    After all, you cannot read a book before reading it.  It would end
    up that a fool like me would first read a section in a printed
    book, and not get hurt, and then look up that section in Info, and
    get hurt.

Yes, you can write a `sandbox' to hold the code executed by Emacs Lisp
expressions in Info, but why bother?  Why add to the cognitive load of
people writing or using Texinfo?  

You would have to explicitly or implicitly mark Emacs Lisp that can be
evaluated in Emacs Info as part of an address@hidden' segment.  (At
the moment, the feature is implicitly marked and disabled by default.)

An address@hidden' segment cannot be output in other formats.  This
feature makes Texinfo further become the kind of language that has
different source forms for different kinds of output.  Texinfo already
suffers from many conditionals -- we already have trouble with authors
who forget that they are writing for plain text as well as graphical
and audio outputs.  I know, since I have read such documents.

So, you have a considerable persuasion job.  You may succeed, but you
must show why the one method is better than alternatives.

    Does not every internet browser have this same risky security

No.  I am puzzled that you should say this.  This statement does not
make any sense to me.  I have never had W3 mode, or W3M mode, or
Mozilla cause me any security trouble that I know of.  Should I be
worried?  Can they automatically execute code that can hurt me without
me consciously giving the command?

    Robert J. Chassell                         Rattlesnake Enterprises
    http://www.rattlesnake.com                  GnuPG Key ID: 004B4AC8
    http://www.teak.cc                             address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]