Hi John, (01)
In the case of a three year old child as you mentioned below, I only got
involved in that particular case after the failure of a software company to
meet contracted delivery terms with the 911 response company re fitness for
use of the software and contract liability. Although most software licenses
work for low cost high volume products like Word, Android or even OSs,
software contract developers are subject to statutory processes based on the
purchaser's perceived expectations. So I don't know what concerns, if any,
were devoted to that issue in the design phase for this particular system.
I didn't get called in until there was an operational problem of software
failure that resulted in a law suit. (02)
But for my suggestion to help get validated ontological columns (i.e., an
NLP-aware program that revalidates newly entered rows in depth using
background processing of the slightly validated string data), I personally
suggest that the operator should immediately forget about the program and
just get the paramedics and police started toward the right location. While
the kid is crying on the phone, the operator might be able to collect some
other information, such as the kid's last name, which might give a location,
but a three year old doesn't even know how to say where the place is
located! So unless the dialING number is traceable back to the physical
address, the kid's mom is in deep kimche. (03)
But in the case of the 911 system, even though license terms are constructed
by the software company's attorneys to be as bullet proof as local laws
allow, there will still be major liability concerns that drive management
(and their insurers) to pay close attention to patents, Sarbanes Oxley, ISO
certification, potential data-responsible damage situations, and other
safety-oriented or liability-limiting measures. (04)
The whole point, at least from my view, is that singularity of the column
semantics is simply impractical to provide, even if theoretically highly
desireable. I agree that singular ontological columns are "better" once you
have them validated, but it is singularly impractical to do so in the
MAJORITY of realistic DB applications, based on my experience. (05)
There is no substitute for having a buncha old row cases to develop a feel
for how data gets entered in actual input samples, especially by each
operator due to their personal data entry idiosyncrasies. My document (the
7,209,923) explains how to collect, organize and analyze those row cases so
that whatever ontological purity happens to exist in the data can be
enhanced by informed, computation-intensive, reinterpretation processes. (06)
But that validation can be complex also, perhaps more so than the DB
application itself. In many web service applications, the data coming in is
validated over the web in real time, especially if a purchase is taking
place. For example, if you give your email address, the system can ping the
IP from which you entered it, and can determine if the person who entered
the data has her email client up and running at the time of data entry.
Similar kinds of validators can be constructed for bank accounts, routing
numbers, zip codes, credit card numbers, SSNs, driver's licenses and so
forth. (07)
But if you know all examples of how a column has been filled out in the
past, due to many, many instances of collected data rows, you have a good
chance of building a NLP routine that (more) accurately discerns INTENDED
meaning expressed by the operator. Since most columns can't be validated by
external network means, that makes for some complicated processing; this
could be more complex than the typical DB application's complexity. (08)
Ultimately, computers and networks will become fast enough to do full scale
validation in real-time during input (maybe by using conversational voice
response and Q&A techniques), but that time is far in the future with
current technology. So for the moment, ontological purity is limited to a
very few percent of columns in real applications needing it, IMHO. (09)
I'm sorry if this message comes across as too strident; I'm not trying to
pop the ontological bubble, just to inform it. One caller even accused me
of "throwing rocks at the ontologists", though in teasing jest! (010)
But I think this point really, really needs to be understood before placing
a lot of faith in assuming ontological purity. It is highly desireable and
highly unlikely to be provided in practice at this time in the technological
rollout sequence. At least, not until I get my prototype running bug free!
Just the savings in insurance premiums and legal expenses will more than
justify the cost for most applications. (011)
HTH,
-Rich (012)
Sincerely,
Rich Cooper
EnglishLogicKernel.com
Rich AT EnglishLogicKernel DOT com
9 4 9 \ 5 2 5 - 5 7 1 2 (013)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F. Sowa
Sent: Tuesday, September 14, 2010 7:51 PM
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Language vs Logic (014)
Rich, (015)
I was about to say that is unbelievable. But unfortunately, knowing
typical computer systems, I find it all too believable: (016)
> In another case, a 911 emergency response system, the desk time required
> to enter a new caller (one who has never phoned that 911 service before)
> is nearly twenty minutes! There is no way that the users can be forced
> to carefully think through every field of data so that they put off
> responding to heart attacks with twenty minute delays to fill out
> database forms. The person at the other end of the phone is often
> emotionally stressed, and not willing to work with the form filling
> operators. (017)
In logic, you can say "For all X, there exists a Y, such that P(X,Y)"
But when your DB has many such constraints, you want to allow many
or even most of the implicit Y's to be left unspecified. (018)
What do you do if the caller is a 3-year-old child, whose mother
explained how to call 911 if she had an epileptic fit? (019)
John (020)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (021)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (022)
|