[PALNet-general] Utilities for analyzing keyboards?
dzo at BISHARAT.NET
Mon Jun 30 13:03:05 UTC 2008
Hi Martin, all,
Martin, your point is well taken, but I'd see it as a complementary
approach, not an alternative. It seems like there are so many possible ways
of providing for extended character and diacritic input, that tools to
analyze them before attempting user trials with a select number would be
very helpful. This is the usual path of product research and development in
other spheres of activity as well.
Re placement (Tunde's suggestion; Andrew's and Martin's replies), I think
there are a lot of people who could suggest better total rearrangements
(remixes?) of the keyboard. However, until the hoped-for day when LED key
keyboards like the Optimus become a lot less expensive, our strategy for
Latin-based orthographies will have to be to work with the base QWERTY or
AZERTY layouts. This is not only for users who are familiar with them, but
also because in effect repainting the keys would lead to a whole other set
of problems and then too, and for total novices to computing, it may be
unfair to initiate them into a totally different key arrangement than one
they may encounter elsewhere (sort of the problem the DVORAK layout has had
From: Martin Benjamin [mailto:piperzen at gmail.com] On Behalf Of Martin
Sent: Monday, June 30, 2008 6:30 AM
To: Andrew Cunningham
Cc: Tunde Adegbola; keyboards at kasahorow.com; 'A12n tech support';
'Indigenous Languages and Technology'
Subject: Re: [PALNet-general] Utilities for analyzing keyboards?
As an alternative approach, what about doing some tests with actual
users? People could be given sample texts to type on a few different
keyboard layouts for a given language. The researchers could (a)
observe the people typing, and (b) ask them for their opinions.
On a personal note, I have been exposed to a number of different
European keyboards recently. The main difficulties I have experienced:
* finding specific characters on the keyboard, such as '
* keys that are too busy (do too many different things depending on alt
and shift combos)
On the other hand, letter placement is not too problematic; although the
z and y keys on the keyboard most commonly used in my locale are
reversed from qwerty, I've found that my brain is able to make the flip
without too much grief. Based on that (limited) experience, I would
suggest that it is easy for people to learn letter placement, but more
difficult for them to memorize complicated escape sequences. How to
confirm this hunch? A little human observation could answer a lot of
questions that a complicated statistical analysis would never address.
Andrew Cunningham wrote:
> On Sun, June 29, 2008 3:52 pm, Tunde Adegbola wrote:
> > One feature that may also be useful is to determine the most frequent
> > pairs of characters so as to be able to arrange the layout for
> > fingers to type (each character of) frequently occurring pairs of
> > character
> > Tunde
> Although such data would be most useful if you ditch the notion of a
> qwerty keyboard and develop a layout from scratch based on single
> character frequency and character pair frequencies.
> A common approach on non-Latin scripts.
> The problem with African languages would be weaning people away from the
> US or French keyboard layouts that they are used to.
> For people who aren't overly familiar with computers, such a change will
> have less impact, but would require developing physical keyboards, with
> the characters printed on the keys in order to facilitate typing.
> Thinking about keyboard layouts, ISO-9995 allows for two keys to access
> level 2 characters (the two shift keys) and dictates their position). It
> also allows for one or more keys to access level 3 (the AltGr key). I
> wonder why most keyboards restrict them selves to a single key for
> accessing level 3 characters, rather than having two such keys to
> facilitate typing? It would make things easier.
> Andrew Cunningham
> Research and Development Coordinator
> State Library of Victoria
> andrewc at vicnet.net.au
> PALNet-general mailing list
> PALNet-general at panafril10n.net
More information about the Ilat