Re: Great SWT Program
nebulous99@gmail.com wrote:
On Nov 22, 5:07 am, b...@pvv.ntnu.no (Bent C Dalager) wrote:
Of course it is: it ensures that the selection [remainder of irrelevancies deleted]
Selection? We're discussing doing a search, not doing a selection.
The example that you re-introduced was one of making a selection using
incremental search. Surely you remember this?
Let us discuss each thing in isolation please. Incremental search,
selection, etc.; -- no more changing the subject in midstream. :P
Furthermore, of course, it is useful while seaching also: You write
letters until you get to the first hit of the word you want, then you
type C-w to complete that word. This way, you can search for the
remaining instances of that word without having to write it all.
How odd and unnatural.
To you.
No, period. Suppose I'm searching for "student" and there's also the
name "Stuart" and the word "stunned" in the document but nothing else
starting with "stu". If I type "stu" and then this C-w binding of
yours, it must either complain that it's ambiguous or pick one of the
three. In the former case there is a 0% chance that it picks "student"
and in the latter there is a 33% chance that it picks "student".
Better than even odds that it's useless. May as well just type a "d"
instead of C-w, and now the only matches are for "student", which is
just what the doctor ordered -- and with an unchorded fourth keystroke
to boot.
If you are trying to find student using incremental search, you would
type it out until it was either a) completely unambiguous, or b) enough
unambiguous that you feel able to quickly navigate using the keyboard
equivalent of find-next/find-previous.
(No doubt you can up the odds from 33% to 50% above by twiddling
something to do a case-sensitive search. Of course that then risks
false negatives, e.g. if an occurrence of "student" is at the start of
a sentence, and hence capitalized as "Student", e.g. in "Students
frequently get this question wrong" or some such sentence.
I've seen incremental search implemenations that default to
case-insensitive, unless they notice mixed-case in your typing. While
this doesn't help when you want to search all lower-case, it DOES help
in most common situations.
This is not a manner of usage that will occur
to most people, and it sounds clunky and awkward even so.
It's not supposed to "occur to most people" - it's supposed to be
learned by the users.
Learned from *where*? This is a technique, which won't be in a
rudimentary command reference. The interface should either make it
easy for a given useful technique to be discovered by a user de novo,
which requires it be an obvious combined application of already-
familiar commands, or else provide explicit documentation or a tip-of-
the-day or something, if users are to know of the technique.
Otherwise, it needs to be reasonably usable without the technique, or
else it becomes dependent on live tutoring to be usable, which is a
showstopping failure in the usability department any time it occurs.
Allow me to make an analogy.
Emacs is a power tool and swiss army knife combo. Notepad is a pen and
paper. Its fair to say that the power tool needs some training, but can
do much more when used by an experienced user.
There are enough people who prefer emacs (or other "text" based editors)
that I feel safe in asserting that its not impossible to learn. It is
definitely *not* a tool for every-man. Its a tool for those of us who
need a little more power out of their editor.
Perhaps the help mechanism could be improved, but that doesn't mean the
features are useless or even clumsy.
--
Daniel Pitts' Tech Blog: <http://virtualinfinity.net/wordpress/>