Re: Info sucks?
On Thu, Sep 17, 1998 at 01:23:27AM -0500, Manoj Srivastava wrote:
> I really think we should not be having this discussion. To
> prove it, read my comments below.
> Avery> But there are good and bad interfaces, and I take that as an
> Avery> absolute fact.
We can say that there are good interfaces and bad interfaces, but we can't
have a discussion about what defines one or the other? How odd. Or do you
insist that there are no universal concepts we can follow that lead to a
good user interface? How defeatist. Or are you saying that if YOU like it,
it must be perfect, so we have no need to discuss improving it? How
I've read the Debian mailing lists for almost a year now, and most people
here bother to think about other people's points before spouting their own.
I have thought about yours: I have given you a list of interface features
that make a program easier to learn and easier to use for people (the vast
majority of the population) that can't read tutorials and remember
everything the first time. For those people (a tremendous portion of the
population), programming in emacs-lisp is simply impossible. Some people (a
massive quantity of them) do not have sufficiently logical minds to be able
to program. We should not require them to.
I have nothing against the emacs interface, and little against the vi
interface :), for power users. I have some major complaints about the UI in
most other Unix programs. But they all ignore fundamental user interface
design concepts that, in many scientific studies, have been _proven_ to make
the vast majority of people more productive.
A wonderfully detailed summary of several UI studies can be found in the
book "The Trouble with Computers" by Thomas K. Landauer. It's a bit dry,
but the information inside is worth the effort.
Please read his results before you tell me again that this discussion is
On the original point of this discussion, he also talks about the
development of a truly efficient book-reader application, and tells how very
difficult it was to make the online book more efficient than a paper one.
But they did it -- and in my opinion, no one ever has since.
And such improvements do NOT, in themselves, make power users less
productive. Don't take away emacs-lisp -- interfaces that are "easy"
through oversimplicity are doomed to failure. Hey, emacs-lisp is great,
though I personally dislike the language. But don't make it _necessary_ to
hack emacs-lisp just so I can change the tab sizes in my c-mode autoindent.
With this, I leave the thread, because I have no more good points to make on