[DECtalk] question about Kurzweil 1000 voices

[Don]

On 6/16/2021 8:57 AM, Piotr Machacz wrote:
> That's not even the only problem with Espeak. The other serious issue which
> makes it impossible to use for many projects is that it's GPL licensed. So even
> if you were to fork it, made a much better sounding version on its basis, the
> GPL license forces you to open-source all of your work as well which makes it
> almost impossible to use for anything commercial.

Actually, I plan on releasing everything as open source.  But, anything
GPL would require me to insist that anyone USING my improvements would
also be burdened with the GPL.

My goal is to encourage my stuff to be used.  The more restrictions I
impose on it -- or have forced on me to impose on others -- the less
likely it is for someone to embrace it and the goals I'm trying to achieve.

And those *goals* are the important thing -- not a copyright on a piece
of code or a patent on an idea!

> So in the case of using it
> with a bluetooth headset or any other hardware even if you got Espeak going on
> it, you'd have to also completely open-source all the firmware for the device
> which could open you up to a lot of serious problems.
>
> THat being said, I too would love to hear what you've been working on. ☺️

There is a belief that you can design an "inclusive" device for no more
cost/effort than you can design a device that excludes users with
varied abilities/constraints.

This is fantasy.  If it was true, then it would be done every day!
You'd not need devices tailored for folks with different sets of
abilities.  You'd be able to take advantage of the economies of
scale (mass production) for EVERY type of user.

Instead, you have portions of the user base that have to be
addressed with SPECIALIZED products.  With smaller markets,
those products tend to have higher costs (cost is what it
takes to MAKE something).  And, often the needs of those
users have higher "support" costs -- it's more expensive
to sell a wheelchair than a motorized skateboard!  You
can walk into a Walmart and buy a motorized board but would
need to be FITTED for a wheelchair!

These conspire to drive the PRICE of the item up (price is
what you charge to SELL something).

This makes the item harder to sell.  Which makes it even
more expensive.

Or, causes the product to be discontinued or the vendor
to exit the market.

So, you REALLY want to be able to sell the same device to
EVERYONE -- instead of burdening one portion of the
marketplace with a more costly version of The Same Product.

One of the things hampering accommodating folks with other
needs is the lack of tools to assist with those sorts of
development efforts.  For example, you see all sorts of
tools that help you draw pretty pictures on a screen...
but nothing that makes it easy for you to express that
same information audibly.  Or tactile.

And, because of that, you don't have developers THINKING
about how they are presenting their "output".  They
spend more time thinking about colors and fonts and
icons than the CONTENT that they are trying to convey.

How many are even aware that many people can't tell red
from green?  So, why pick red and green to signify stop
and go?  Or, that many can't READ what's written on
the screen due to vision problems?  Or, that those vision
problems are different in, for example, folks with a
congenital vision problem (like blind from birth) vs. those
who have lost their vision from an accident or diabetic
retinopathy.  Or, those in the process of losing their
vision to macular degeneration?

In other words, a braille display would be useless for
grandma losing her vision to cataracts -- she's unlikely
to ever invest the time in learning braille!

Similarly, at advanced ages, how good would her hearing be?

And, how do you convert a picture into words?

How does a sighted paraplegic type on your keyboard?

If you truly want something to be universally accessible,
then you have to consider each of these potential users
in the design of your user interface.  You can't just
"bolt on" some sort of adapter (like a screen reader)
and consider the problem solved!

Speech is just one "output modality" that addresses some
*portion* of the user base.  If you need a PC in order to
create it, then you're now saddled with the cost and size
of a PC before you even get started!