Matthew said:
You should design whatever you believe is going to be most
profitable.
In a situation so subject to personal prejudices shouldn't decision
making be based on something more than belief? To know that something
will be more profitable, to be in a position to demonstrate a cause and
effect relationship, would be grounds for applying that knowledge to a
design decision.
But the knowledge on which these sorts of decisions could be based
doesn't seem to exist. At least nobody engaging in these debates on
Usenet seems to be in a position to direct anyone to anything
substantive (with the exception of the people questioning the
possibility of gating meaningful statistics from the Internet as they
can substantiate the technical characteristics of HTTP that deny the
possibility of gathering accurate statistics).
I don't care if it's browser, resolution, flash, js or
what.... You need to look at the income potential of what
you've got vs. using other methods.
If I do something that totally alienates 5% of my viewers but
increases sales by 10% for the rest, that's obviously a step
forward.
That is a net gain of 4.5%. Which makes it very important to be certain
that this change will deliver a 10% increase in sales, as if it turns
out to be only a 5% increase the net change is -0.25%.
Of course if you can make this (unspecified) change without alienating
that 5% then it isn't so important if it only results in a 5% increase
in sails for the remaining 95% as it is no longer trying to offset a
loss in sales.
That last point seems to make it very important to be able to identify
what it is about these changes that will result in this 10% increase in
sales. So that you can be certain that it cannot be achieved without
consequential alienation of potential customers.
From you list, take javascript (as it is a particular interest of mine);
there is certainly a belief that a level of interactivity facilitated
with javascript somehow increases the profitability (or in some other
sense "value") of a web site. Specific details of this relationship are
somewhat thin on the ground, but for the sake of argument lets say it is
true.
It is certainly true that a web site that is dependent on javascript to
be useable will alienate a percentage of its potential visitors (all of
those using javascript disabled/incapable browsers plus any using script
enabled browsers that do not support the required features).
It is implicit in the decision to create a javascript dependent web site
that the benefits gained from the interactivity facilitated by
client-side scripting are believed to outweigh the losses resulting from
the alienation of the script disabled/incapable (assuming that the
dependency results from a decision rather than just
ignorance/incompetence).
But those benefits must follow from what the script actually does, not
from the fact that the site is dependent on javascript, or from the
exclusion of the javascript disabled/incapable. (The users who are going
to be responsible for the expected increase in sales will be unaware of
the javascript dependency and the exclusion of the javascript
disabled/incapable so they cannot be influenced by them in their
spending decisions.)
A specific example of javascript facilitated interactivity might be the
pop-up navigation menu. These are often implemented with the
data/text/URLs needed defined in javascript structures and the result is
totally javascript depended, and imposes that dependency on the entire
site as uses without javascript find themselves with no means of
navigating the site.
As I said, the implication of the decision to use a pop-up menu
implemented in this way is that the benefits it will bring outweigh the
losses that will follow from the exclusion of the javascript
disabled/incapable.
Another pop-up menu implementation might be designed to take advantage
of javascript's ability to manipulate HTML. In this case the data needed
for the menu would be defined as, say, nested lists (UL) of links, and
the javascript would take those lists and modify their display
characteristics, positions and associated event handlers to transform
them into the pop-up menu. Once that transformation had happened the
result could be indistinguishable form the previous javascript dependent
implementation, and so any benefits following from the preceding menu
implementation must also follow from this one.
However, this design has not excluded the javascript disabled because
the failure of the script to act has resulted in the nested lists of
links remaining in the HTML, where the users of script
disabled/incapable browsers may employ them directly to navigate the
site.
So for a pop-up menu, any benefits following form its use are available
to anyone who was going to see them anyway, but the alienation hangs on
a script implementation decision and don't follow from the use of
javascript as such.
It tuns out that the vast bulk of the interactivity that can be
facilitated with javascript is quite amenable to implementation in a way
that is layered on top of HTML and/or manipulates it without any need to
introduce any dependency on scripting for the bottom line functionality
of a web site. Modern desktop browsers (the ones "targeted" by the
people who argue that 90% is good enough) facilitate so much content
manipulation that virtually anything is possible. So in most cases the
decision to use a script implementation that would alienate a percentage
of potential users is an active decision to reduce potential sales, as
any perceived gain could be achieved without it through the use of a
better designed script implementation.
And the same goes for the other items in your list. If there is a
benefit in designing for a fixed (expected/standard/common) width, and
CSS can give you the same presentation at that width, but in a way that
is fluid at other widths, you have all that could be gained from the
fixed width design and no penalty. Flash can be a bit more complex but
it certainly isn't beyond facilitation reasonable degradation for the
users of browsers without the required plug-in.
It seems to me that if something is going to be traded-off, some benefit
for some (but smaller) loss, that it is important to know that the
relationship between the two is real (that the benefit cannot be fully
(or mostly) achieved without the loss), and that the benefit will
outweigh the loss.
Richard.
