interactive keys and others

[Mike Dallwitz]

Mary Barkworth wrote:

> I use [interactive keys] only as a last resort. This may be because I like
> reading and find that I absorb information through reading, which I enjoy,
> and I like reading.

OK, if your goal is to absorb information. But if your goal is accurate
identification, interactive keys should not be a last resort.

> I also like seeing what comes out around my plant. Sure, I could find this
> out with Intkey or other programs - and get information designed for a
> computer.

What comes out around the plant is not of much significance unless the key
is specifically designed to reflect the classification, which usually makes
it less effective for identification. In any case, Intkey automatically
gives this information, in the form of a list (designed for people!) of
remaining and eliminated taxa, together with the number of differences from
the selected attributes.

Phenetic (and cladistic) trees, and lists of nearest neighbours are not
directly available in Intkey, but they are readily available via other
programs in the DELTA suite, and could be made available if required via
general mechanisms for accessing 'external' information in Intkey. (In fact,
list of nearest neighbours were supplied in early versions of Les Watson's
Grass Genera database.)

> There is a reason publishers spend time and effort choosing fonts, layout
> etc.

Books and journals are usually printed from PDF files nowadays. These (or
files derived from the same sources) could made available in Intkey. Of
course, there are still some advantages in reading them on paper. But there
are also advantages in having them on a computer, e.g. searching.
Incidentally, screens that pivot to portrait orientation are now quite
cheaply available, and make it much easier to read typical PDF files
designed for printing.

We've taken considerable trouble to obtain fairly readable output in Intkey.
For example, look at the output of the 'differences' option. The output is
indented to improve readability, words that are needed in other contexts but
are unnecessary here are omitted, and bold font is used to highlight the
most significant parts.

> For analyses, yes, a database of information is great.

And also for identification :-).

> Humans are not computers.

Right. But computers are better than humans at some things, so it makes
sense to use them for those things. You don't see many people doing
complicated arithmetic on pieces of paper nowadays.

> Also, customs officials, taxonomists, and general public have different
> motivations for identification. This also plays into which mode they prefer
> to use. One size does not fit all.

They may have different motivations, but they have at least some goals in
common: convenience, speed, accuracy, robustness. I think that one piece of
identification software (not necessarily one currently in existence) _can_
fit all. It's the data, particularly the characters, that have to be
tailored to particular audiences (as you mention in a later posting).

--
Mike Dallwitz
Contact information: http://delta-intkey.com/contact/dallwitz.htm
DELTA home page: http://delta-intkey.com