On programmatically achieving abstract optical effects with an eye to beauty
and design standards (warning: huge philosophical digression)…

On 6 June 2011 07:48, Alan Gresley <[email protected]> wrote:

> I guess you're learning.....


Not to start qualitative debates on CSS lists? Veeeery slowly. People giving
me the time of day certainly isn't helping ;)


 I thought inset and outset borders were in CSS2. Why are you cynical of
> CSS3?


Inset and outset borders are a great example of an abstract optical effect
getting a programmatic definition in CSS2. CSS3 offers many more
opportunities (border-image, double border, gradients, box shadow, etc). I
was basically being a design snob, inferring that these were a slightly less
dramatic analogy of giving Comic Sans and Papyrus fonts to somebody making a
business presentation. The properties described can be created using
long-established alternative means by someone sufficiently dedicated to use
graphics software and experiment with the desired visual outcome regardless
of what can quickly be written in CSS. With specific regards to plain old
CSS2 borders, I would test and adjust each combination of colours
individually.


So if you haven't tested, how is it that you can express such an opinion?
>

I've done plenty of testing, but mostly with the purpose of achieving
designs to a certain standard for extensive public-facing sites, as opposed
to arbitrary quantitative feature-detection (ie does the inset border style
exist? check; can we use a variety of colours? check; problem solved).
Basically it comes down to your ambition of vision and extensible
replication with regards to artistic control. This is not to say I won't use
purely ornamental CSS3 extensions at all, but they are abused by programmers
too lazy to create the same effects to a better standard without said
extensions.


Can you prove that absolutely?
>
>
This is the crux of the matter: there's no automated unit test for aesthetic
value and optical illusion effectiveness.



>  In short, when you're
>> talking about optical illusions and a certain level of detail with
>> coherent
>> aesthetics, I don't believe it's possible to algorithmically generate the
>> lot based on small input values.
>>
>
>
> I don't follow what you are saying here.


Basically, while the CSS your provided effectively achieves an optical
illusion that suggests an embossed panel by virtue of consistent lighter and
darker geometric shapes suggesting a light source over a 3 dimensional
object, I don't believe the 21st century man on the street would ever
consider this worth representing as a standalone 'feature'. In the early 90s
it might've invoked a sense of cool because it implied the upper limits of
popular emerging computer graphics interfaces of the time. I'd struggle to
find a client who would take me seriously if I proposed this as a design.


Yes, I have. See the example above and the example in my reply to Andrew.
> I'm still working on the maths but I believe I have hit the most aesthetic
> values.
>

Again, I'm very intrigued in the notion of 'working on the maths' to get the
'most aesthetic values', especially seeing as you believe that target has
already been achieved by a 4-tone blue box whose inset panel has a higher
hue saturation than the outer surface. But it is all relative — it *is*
possible to achieve a certain standard of raised box effect based on
procedurally generated code. It's just that in the world I inhabit, that
standard isn't remotely good enough for any serious application — to the
point where it's a bad idea to entertain it.

All sniping aside, I'm genuinely interested in how far people can get with
this exercise precisely because from where I come from it's a false idol —
but it's the insistent and the adventurous that set the standards after all
and I'm keen on following any technical developments on this thread…


Regards,
Barney Carroll

[email protected]
07594 506 381
______________________________________________________________________
css-discuss [[email protected]]
http://www.css-discuss.org/mailman/listinfo/css-d
List wiki/FAQ -- http://css-discuss.incutio.com/
List policies -- http://css-discuss.org/policies.html
Supported by evolt.org -- http://www.evolt.org/help_support_evolt/

Reply via email to