sxemacs-devel
[Top] [All Lists]

Re: On usability (draft)

From: Steve Youngs <steve@xxxxxxxxxxx>
Subject: Re: On usability (draft)
Date: Tue, 28 Dec 2004 12:51:12 +1000
Organization: The SXEmacs Project
User-agent: Gnus/5.110003 (No Gnus v0.3) sxemacs/22.1
* Johann Oskarsson <Johann> writes:

  > I think I'll take the time to write a bit on usability, becouse the
  > term "user friendly" is highly misunderstood.  I think we can safely
  > blame Microsoft for that.

I think we can safely blame M$ for a lot of things.

  > The term "user friendly" is supposed to mean "an interface on a
  > device, for use by humans, that makes it's usage easy" and that's
  > the definition I use.

Me too, except that I tack on a "and/or intuitive for those few humans
that possess common sense".

  > For emacs, it means "text editing expert friendly".

I don't totally agree with that, but in the light of not being able to
come up with something better right now, I'll go along with it (for
the time being).

  > Yes, the learnign curve for emacs is steep, but once there, it's
  > quite intuitive.

Agree.

  > I've been able to guess behaviour many people on #emacs are (or
  > were) quite unaware of.  (I'm talking about C-u M-x unload-feature
  > vs. M-x unload-feature -- yes, I correctly *guessed* the meaning
  > of C-u before I even read the docs)

Thats terrific!  Except for one thing... I've looked at the doc string
for `unload-feature' and I've looked at the code for `unload-feature'
(I was curious because I've never heard of using C-u on it)... and it
_doesn't_ use a prefix arg anywhere.  So doing C-u M-x unload-feature
would be no different from M-x unload-feature.

But!

I consider that a bug.  There is no way to invoke unload-feature
interactively (via M-x) and use the FORCE argument.  (think I'll send
the XE folks a patch)

  > But there are still many "usability issues" to fix.  And I think
  > most of them fall under "can (sx)emacs do X without writing lisp?"

Without the _user_ needing to write lisp, correct?  In other words,
we'd just write it for them.  We might want to keep an eye on bloat
here. 

  > Not all users know or care to learn lisp -- I didn't until I
  > needed my window number mode.  Until I made it, more experienced
  > lispers didn't think it could be done.  

Or didn't think it was all that useful, or didn't think you were going
about it the right way.[1]

  > M-x customize helps, but it goes only so far.

The custom interface is a big fat hairy ugly hack.  It works for most
things most of the time, but as you say, it only goes _so_ far.  IMO
we should come up with something better.

  > There are lots of issues that makes emacs slow and sometimes
  > bloaty, we need to find these defects, order them (imo) according
  > to "usability severity" and fix them in that order.

Good.  I like it.

  > One such issue is the lack of features to enable gnus (or any
  > other mail reader) to create efficient mbox handler.  There is no
  > need to keep the entire mbox in memory,

And this is one of the reasons we have the nnml backend.

  > the lack of low-level file handling in elisp.

Yes, this would be good to address.

  > There's more I wanted to say, but I'll leave it to later.  I mean
  > this as a draft for an article to appear on the website soon next
  > year.

Good stuff!

  > There was some stuff I want to respond to Steve's comments but I'll
  > leave that to later too.  So *snip* ;)

aww, I feel all dejected now. :-(

:-P


Footnotes: 
[1]  Some of me still doesn't. :-P

-- 
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
|                        In space,                         |
|             No one can hear you rip a stinky             |
|------------------------------------<steve@xxxxxxxxxxx>---|

Attachment: pgpNdiKqj8YbI.pgp
Description: PGP signature

<Prev in Thread] Current Thread [Next in Thread>