Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)
John Doe wrote:
> David Maynard <nospam@private.net> wrote:
>
>>John Doe wrote:
>>
>>>Mxsmanic <mxsmanic@hotmail.com> wrote:
>>>
>>>>John Doe writes:
>
>
>>>>>Speech recognition can be extremely frustrating, but sooner or
>>>>>later it will be usable by most people.
>>>>
>>>>Speech recognition, like a GUI, is only a solution to certain
>>>>problems, not all.
>>>
>>>
>>>It's a user interface. It's a way to operate a computer. Highly
>>>likely it will be the way most computers and computer-based
>>>machines will be operated in the future.
>>>
>>>
>>>
>>>>I can type faster and more accurately than I can speak.
>>>
>>>
>>>That is very unusual.
>>
>>Not really.
>
>
> In your opinion.
>
>
>>You're probably thinking of 'text' vs speaking the same text.
>
>
> I'm thinking of what Mxsmanic said.
>
>
>>He's being rather subtle
>
>
> Doonesbury includes some comedy based on ideas like that. A
> Chinese dictator was so disabled that no one could understand what
> he was saying. His aid was running the country.
>
>
>>about it but the point he's making is that computers can be used
>>for lots of things and not all of them are simple text nor
>>things that lend themselves to easy verbalization.
>
>
> My reply was based on what Mxsmanic said.
>
> Now we will argue what you want to argue. At least while there is
> any point to it.
>
>
>>For example, this linux command line plot FastE packet rate vs
>>packet size:
>>
>>echo 'pad=20; plot [64:1518] (100*10**6)/((pad+x)*8)' | gnuplot
>>-persist Now you and I might not be 'quick' to create that
>>command but those skilled in the art are and it doesn't lend
>>itself to verbalization very well
>
>
> Speech recognition doesn't mean that characters have to be
> enunciated one at a time.
>
>
>>(unless you're simply reading it back but that begs the issue of
>>typing it to begin with in order to read it back),
>
>
> Typing it to begin with in order to read it back?
>
>
>>and a lot of complex math doesn't. Neither do compiler command
>>lines.
>
>
> I do not know what you're trying to say.
That's obvious. Might have helped if you'd have tried reading it instead of
looking to find line by line 'retorts', but no matter.
Even in human to human interaction it's often better to write things out,
or put it on the blackboard, rather that verbalization because some things
just don't work as well in pure speech form. But I don't want to interfere
with your illusions about your speech recognition toy so that's the end of
my efforts in this regard.
>>Now, some day we might be able to tell the computer to "plot
>>FastE packet rate vs packet size for me" but that's way past
>>simple 'voice recognition' and gets into advanced artificial
>>intelligence.
>
>
> That isn't necessary for verbal control.
No one said it was but since you didn't bother with reading comprehension
you wouldn't notice.
> Speech control does not
> necessarily mean the computer has to understand like a human
> understands. The objective is to do what you could do by hand, and
> there are many ways to efficiently enable that. A noticeable
> difference is like not having to enter keystrokes on a command
> line, even though the computer might receive the input as if you
> did.
>
> What experience do you have with speech recognition?
About 30 years worth, off and on, as part of my work in artificial
intelligence.
> What
> experience do have with entering/processing/outputting ordinary
> keystrokes?
Oh, none at all. My machines read my mind.
John Doe wrote:
> David Maynard <nospam@private.net> wrote:
>
>>John Doe wrote:
>>
>>>Mxsmanic <mxsmanic@hotmail.com> wrote:
>>>
>>>>John Doe writes:
>
>
>>>>>Speech recognition can be extremely frustrating, but sooner or
>>>>>later it will be usable by most people.
>>>>
>>>>Speech recognition, like a GUI, is only a solution to certain
>>>>problems, not all.
>>>
>>>
>>>It's a user interface. It's a way to operate a computer. Highly
>>>likely it will be the way most computers and computer-based
>>>machines will be operated in the future.
>>>
>>>
>>>
>>>>I can type faster and more accurately than I can speak.
>>>
>>>
>>>That is very unusual.
>>
>>Not really.
>
>
> In your opinion.
>
>
>>You're probably thinking of 'text' vs speaking the same text.
>
>
> I'm thinking of what Mxsmanic said.
>
>
>>He's being rather subtle
>
>
> Doonesbury includes some comedy based on ideas like that. A
> Chinese dictator was so disabled that no one could understand what
> he was saying. His aid was running the country.
>
>
>>about it but the point he's making is that computers can be used
>>for lots of things and not all of them are simple text nor
>>things that lend themselves to easy verbalization.
>
>
> My reply was based on what Mxsmanic said.
>
> Now we will argue what you want to argue. At least while there is
> any point to it.
>
>
>>For example, this linux command line plot FastE packet rate vs
>>packet size:
>>
>>echo 'pad=20; plot [64:1518] (100*10**6)/((pad+x)*8)' | gnuplot
>>-persist Now you and I might not be 'quick' to create that
>>command but those skilled in the art are and it doesn't lend
>>itself to verbalization very well
>
>
> Speech recognition doesn't mean that characters have to be
> enunciated one at a time.
>
>
>>(unless you're simply reading it back but that begs the issue of
>>typing it to begin with in order to read it back),
>
>
> Typing it to begin with in order to read it back?
>
>
>>and a lot of complex math doesn't. Neither do compiler command
>>lines.
>
>
> I do not know what you're trying to say.
That's obvious. Might have helped if you'd have tried reading it instead of
looking to find line by line 'retorts', but no matter.
Even in human to human interaction it's often better to write things out,
or put it on the blackboard, rather that verbalization because some things
just don't work as well in pure speech form. But I don't want to interfere
with your illusions about your speech recognition toy so that's the end of
my efforts in this regard.
>>Now, some day we might be able to tell the computer to "plot
>>FastE packet rate vs packet size for me" but that's way past
>>simple 'voice recognition' and gets into advanced artificial
>>intelligence.
>
>
> That isn't necessary for verbal control.
No one said it was but since you didn't bother with reading comprehension
you wouldn't notice.
> Speech control does not
> necessarily mean the computer has to understand like a human
> understands. The objective is to do what you could do by hand, and
> there are many ways to efficiently enable that. A noticeable
> difference is like not having to enter keystrokes on a command
> line, even though the computer might receive the input as if you
> did.
>
> What experience do you have with speech recognition?
About 30 years worth, off and on, as part of my work in artificial
intelligence.
> What
> experience do have with entering/processing/outputting ordinary
> keystrokes?
Oh, none at all. My machines read my mind.
Facebook
Twitter
Instagram