Hi Dan,

I've come to the conclusion that "visual language" is for the most
part contradiction in terms.

When I look at LabVIEW, Max or PD, I see an array of iconography and
symbols which are clearly text.  If you replace the word 'or' with a
triangle, you're just replacing a string of symbols with a single
symbol.  This has major drawbacks as there isn't a button on the
standard keyboard for the triangle symbol.

Artists often say they like Max and PD because they are visual.
However you can put boxes where you like, it makes no difference
because visual layout in these languages is entirely secondary syntax.
 Max users often counter this point by saying right-left order gives
execution order, which is an odd point as this is true of almost all
'non-visual' programming languages.

There seems to be an idea in computer interfaces that vision is
somehow 'more advanced' than words. This seems somewhat in denial of
the history of our species.

Myers (Taxonomies of visual programming and program visualization;
1990) defines visual programming as "any system that allows the user
to specify a program in a two (or more) dimensional fashion."  This is
odd as in almost all visual programming languages I've seen, 2D
arrangement has no syntactical value.  However a number of `textual'
languages, notably Haskell and Python, *do* have 2D syntax.  So maybe
they're the answer for your request for visual languages with keyboard
input! :)

In opposition to Myers, I think the value of so-called `visual
languages' is that they are not constrained by visible dimensions *at
all*.  As you are unconstrained by the usual visual property of
adjacency, you can define graphs of arbitrary high dimension,
hypercubes and up.

Considering the Cognitive Dimensions literature it seems clear that
rather than a visual / non-visual language dichotomy, there is instead
a dimension of linguistic/visual integration.  Source code always has
a visual component, usually not part of the syntax but still very
important to communicate conceptual aspects of the program in
secondary syntax.  I think it *is* possible to integrate visuospatial
and linguistic notations more, and research around Dual Coding theory
seems to suggest that the result would be computer languages more
suitable for femine than masculine tendencies.  I am working on a
programming language inspired by the ReacTable for example, that I
hope to demonstrate at the PPIG WIP in Sheffield.

So in conclusion, in my view visual languages are not visual, and that
the dichotomy has been drawn in the wrong place. Instead language and
imagery should be viewed as different aspects of notation and thought,
and we should seek to integrate them in interesting ways, not think of
them in opposition.



On 8 March 2011 09:40, Dan Stowell <dan.stow...@eecs.qmul.ac.uk> wrote:
> Hi,
> I'm fairly new to the PP literature. Thomas's self-promotion led me to his
> interesting 1996 paper on visual programming languages. "Overall, we believe
> that in many respects VPLs offer substantial gains over conventional textual
> languages, but at present their HCI aspects are still under-developed.
> Improvements in secondary notation, in editing, and in searching will
> greatly raise their overall usability."
> As someone who does a lot of work in textual languages - in particular,
> livecoding - I'd be interested to know what the state of the literature is
> on VPLs, especially in comparison against text languages. I haven't been
> able to find a recent survey, any recommendations?
> (Also, as someone who has had RSI, I wonder about the accessibility of VPLs
> via purely keyboard control...)
> Best
> Dan
> On 01/03/2011 17:32, Thomas Green wrote:
>> Er, a spot of self-promotion here ..... the various types of comparison
>> I did in the past led to a framework which attempts to make some sense
>> of the underlying trade-offs, the cognitive dimensions framework,
>> developed by me and lots of other people. Stefano, if you simply want to
>> know whether your new tool works, then you probably just need to do an
>> experiment and stop; but if you want to know why it works (or doesn't),
>> you might take a look at that framework. There's a resources page here:
>> http://www.cl.cam.ac.uk/~afb21/CognitiveDimensions/
>> CDs analysis is quite quick, though very vague. It's actually quite
>> possible that it would reveal problems you've overlooked .....
>> Thomas
>> --
>> The Open University is incorporated by Royal Charter (RC 000391), an
>> exempt charity in England & Wales and a charity registered in Scotland
>> (SC 038302).
> --
> Dan Stowell
> Postdoctoral Research Assistant
> Centre for Digital Music
> Queen Mary, University of London
> Mile End Road, London E1 4NS
> http://www.elec.qmul.ac.uk/digitalmusic/people/dans.htm
> http://www.mcld.co.uk/


Reply via email to