[A12n-Collab] Utilities for analyzing keyboards?

[Don Osborn]

Thanks Walter, This is helpful. Does anyone have any feedback on the format?


It would be excellent to have something practical come out of this. I know
that at least a few others are interested in this issue.

Don


-----Original Message-----
From: a12n-collaboration-bounces  On Behalf Of Walter Bender
Sent: Saturday, June 28, 2008 3:12 AM
To: A12n tech support
Cc: keyboards at kasahorow.com; Indigenous Languages and Technology
Subject: Re: [A12n-Collab] Utilities for analyzing keyboards?

#2 is really nothing more than running the text through a look-up
table, since keyboards don't typically maintain state. Should be
pretty trivial. If we could agree on a format, I could write a script
to process X Windows symbol files.

-walter

On Fri, Jun 27, 2008 at 1:58 PM, Don Osborn <dzo at bisharat.net> wrote:
> In discussing layouts for African languages whose orthographies include
characters or diacritic combinations not supported by American and European
keyboards,* and being aware not only of the multiplicity of possible
layouts, but that various layouts have already been created and some number
of them are used to varying extents, I'm wondering about the potential value
of 2 kinds of utilities to analyze layouts statistically:
>
> 1) A way of "censusing" character placements on keys on existing layouts -
in other words a way to enter keyboard layouts into the utility and have it
count what is assigned to what key, and show what keys have been used for a
given character. It seems to me that this information might be useful in
getting an idea of whether there are some emerging consensuses (or people
for whatever reason thinking in the same lines) regarding placement and
arrangements. Such data might possibly point to things we hadn't realized or
imagined.
>
> 2) Even more useful might be a utility to analyze the work of typing with
different existing and proposed layouts. Say you had alternative layouts for
Yoruba (or any language). The idea would be a way to have a utility in which
you could enter the layout information and then put in a text and figure out
how many keystrokes it takes to enter that text with the alternative
layouts. In principle, once such a thing existed you could also analyze what
fingers are doing the work, but the main idea I have of being able to say
how "efficient" a given layout is in terms of keystroke count for sample
(and hopefully representative) texts, and so anticipate how it might be
received by users.
>
> IMO, one really helpful outcome of any keyboard project at this point in
time would be the ability to systematically analyze what we are proposing
and thus take the learning and quality of production up a notch. Perhaps the
concept of such utilities to do this would interest multiple projects, such
as PanAfrican Localisation Network, AfLaT, OLPC, .... And perhaps commercial
ventures would be interested.
>
> I'll cc the ILAT list since some indigenous languages with extended Latin
character sets presumably have similar issues.
>
> Don Osborn
> Bisharat.net
>
>
> * Mainly "category 3 & 4" orthographies in the system I proposed last
December on A12n-collab:
>
http://lists.kabissa.org/lists/archives/public/a12n-collaboration/msg01059.h
tml
>
>
>
> _______________________________________________
> A12n-collaboration mailing list
> A12n-collaboration bisharat.net
> http://lists.kabissa.org/mailman/listinfo/a12n-collaboration
>
>
_______________________________________________