Re: [ontolog-forum] doing standards [was - Re: Webby objects]

[Amanda Vizedom <amanda.vizedom@xxxxxxxxx>]

William,

I, at least, am not talking about an ontology without any NL involved. I'm talking only about the advantages, in at least some cases, of moving (the illusion of) NL out of the unique-concept-names. In fact, I think that NL rich and accurate enough to be used for good generation, visualization, parsing, disambiguation, indexing, retrieval, and all those great things is missing and wanted in many ontologies. 

I say semi-NL, because to make concept names unique, somewhat disambiguating to core users, and syntactically acceptable, NL-based concept names end up somewhat artificial anyway, not exactly matching NL usage.

I say (the illusion of) NL, because what's captured in these names isn't lexically rich enough or accurate enough to support those great ontology uses I mentioned above. 

Concept names can't really do the work of supplying this semantic assignment role. One reason is the complexity, contextuality, and dynamic nature of how linguistic reference and usage really work. Another reason is the tension between the requirements on names - unique, short and manageable in developer tools, unambiguous -- and the nature of NL expressions of a concepts. 

So, the question for me is not NL vs no NL.  It's how the NL should be incorporated into the ontology. I think that in most cases, if you have diverse users (effectively different NL communities, over language, dialect, domain, role, *or time*), the NL aspect of the ontology needs to be handled in a way that supports processing according to NL principles. It also needs to support things like localization and change (don't want to change IDs because usage changes or new users step in - change/extend the lexical assertions as well. 

In short, the argument is not for less NL, it is for more and better. I say this as an ontologist who is *not* a  linguist.  I've worked on non-toy applications with significant NL mapping requirements, at significant scale. In the process, I've learned a lot about lexicalization of ontology concepts (that is, really, incorporating assertions that map concepts to lexical items) and what the requirements are for doing that in a way that supports good generation, visualization, parsing, disambiguation, indexing, retrieval, and all those great things.  The attempt to load the NL into the concept *names*, rather than using mapping assertions (label, usefor, or otherwise), doesn't get you a whole lot of actual semantics and does get you extra problems.  

Best, 

Amanda

On Wed, Nov 28, 2012 at 8:06 AM, William Frank <williamf.frank@xxxxxxxxx> wrote:

I do not understand how, if machine ontologies are nothing but formal relations between expressions, they have any semantic content?
If the 'ontology' is just a formal system, then it could not be used for any pariticular purpose, unitl it is *assigned* an interpretation outside itself.

For example, suppose there was an ontology with six primtive terms, p1, ,,, p6, and 6 derived terms, d1, d6. d1, for example, is defined to be an entity consisting of exactly one of p1 to p6.

What is this ontology about?  Could it not be some subset of inorganic chemistry, or equally a parts explosion of a carborator?  How do we know what it is intended for unless it is attached to experience outside the system?

For example, John Sowa talks about how chemists use NL expressions like 'carbon dioxide' in very precise ways.  But there is more to organic chemistry than the interlocking of terms; it is mostly about  making observations, doing things in labs and predicting behaviors of real things.  

So, without any NL directly involved, how can even a machine do anything useful with a machine readable ontology?

 

On Wed, Nov 28, 2012 at 1:16 AM, doug foxvog <doug@xxxxxxxxxx> wrote:

On Tue, November 27, 2012 23:26, Patrick Cassidy wrote:
> Using human-friendly terms should not be a problem, if there is even a
> minimum of common collegiality in the process.

Here, the difference is between "should not be" and "is".  The problem
is that human-friendly terms are too friendly -- so that even ontologists
who should know better will think, "i've seen that before -- i know what
it means" and then operate on the meaning that they tho ught they
remembered, instead of checking what the meaning actually is when
they re-use a term after not using it for a while -- or when they actually
have not seen it before, but they think so, because it sounds friendly.

I've seen this happen often with "human-friendly" terms.

> Whenever there is a dispute about whether some term
> should label one concept or another

A big issue arises when there is no dispute -- people just assume
different meanings.  They are unaware that there is a conflict.

One issue is who should actually see the ontology term IDs?  People
don't argue over what is the best name of a method in a Java package.
And most programmers don't willy-nilly use terms from such packages
without taking time to understand what the methods do.

Why should anyone other than ontologists see the ontology term IDs?
[flack jacket on.  8)#  ]

> in one particular ontology, that bare term should be left unused and more
> specific variants can be used for concept representations that have
> different definitions.

Banning all one-word terms would certainly help.

> So, instead of, e.g. a term "Process" (never
> defined the same way in any two upper ontologies I have seen), we might
> have "ContinuousProcess"  for phenomena describable by differential
> equations, or "DiscontinuousProcess"  for an Event represented as a
> series of steps (which may be considered instantaneous, or take some
> finite time, which distinction may lead to more refinement or expansion of
> the labels). When more than one ontology Is being considered, the formal
> way of doing this is just to use namespace prefixes so everyone can use
> the same label, but specify the namespace when creating the logical
> specification.



> I personally really hate using numbers or gensyms for labels,

At one level, i do too!

> as I see no good reason not to use meaningful linguistic labels,

However, i have found that others are not as careful with the
meanings of terms as i think i am.

> and would rather not
> have to create and refer to a "label"  property to remind me what the
> entity is intended to represent.

Such a "label", just like a "user-friendly name", is just a hint as to the
meaning.

> The actual logical specification is of course the definitive
> meaning, but it would be a waste of time to have to
> read that every time one encountered an ontological entity.

But how can one ensure that it is carefully read *the first* time
that any person encounters the entity?

> Linguistic labels can work just fine, used appropriately.

I find that ensuring appropriate use can be a problem.
An ontology system with strong type checking and narrow type
constraints on relations (not concluding the types) can catch
much misuse of terms.

I remember that a few years back, the Cell Line Ontology had
a number of cell types that were subclasses of both Plant Cell
and Animal Cell even though they used numeric IDs and English
labels.  For example, the term labeled Epidermal Cell had both
plant epidermal cells and animal epidermal cells as declared
subtypes.  However, someone at some point (presumably not
a botanist) asserted that Epidermal Cell is a subtype of Animal
Cell -- their system accepted the assertion -- and it was part
of the released ontology.

Not only was the linguistic label property not used appropriately,
the full text comment did not make it crystal clear that epidermal
cells of both plant and animal cells were included in the concept.

-- doug foxvog

> Pat
>
>
>
> Patrick Cassidy
>
> MICRA Inc.
>
> cassidy@xxxxxxxxx
>
> 908-561-3416
>
>
>
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Ed Barkmeyer
> Sent: Tuesday, November 27, 2012 6:52 PM
> To: [ontolog-forum]
> Subject: Re: [ontolog-forum] doing standards [was - Re: Webby objects]
>
>
>
>
>
> Amanda Vizedom wrote:
>
> Yep. I have seen the same thing elsewhere, as well; even though
> ontologists *know* that the concept names (vs. labels) are meaningless to
> the machine, the ontologists are still human language users and slip into
> various errors such as expecting names to carry a content burden and
> fighting over names.  I have also experienced the significant gains in
> usability and efficiency that can come from using concept IDs that are not
> easily interpreted by humans (e.g., hexadecimals at Convera). IMHO, that
> is the way to go -- it's amazing how much confusion is avoided when names
> are obviously there as unique handles but carry no meaning, and N-N
> mapping between NL terms and ontology terms is practiced and supported.
>
>
> Yes, if you can get the community to maintain the discipline.
>
> The other side of that coin is that the magic numbers become the terms of
> the language spoken by the in-crowd, and it becomes a jargon that freezes
> out those with an interest in the domain but lack that time or patience to
> deal with the black speech.  Supply-chain software people know what a
> DELJIT and an IFTMAN are.  They are the codes for EDI message types
> (shipment notifications and freight manifests) that have formal titles,
> aliases, and relatively clear textual descriptions, but the community
> doesn't use the titles or the descriptions.  They pronounce the codes,
> which mean nothing to the accountant or the man on the dock.
>
> -Ed
>
>
>
>
>
>
> Best,
>
> Amanda
>
>
>
> On Tue, Nov 27, 2012 at 5:22 PM, doug foxvog <doug@xxxxxxxxxx> wrote:
>
> On Thu, November 22, 2012 16:59, Obrst, Leo J. wrote:
>> Sure, Amanda, and that's why I (and we) advocate using natural
>> language vocabularies that are linked/mapped to ontologies.
>
> If you are referring to advocating the separation of NL vocabularies from
> ontology term names, we certainly agree.  The ontology needs to express
> an N-N mapping between NL terms and ontology terms.
>
>
>> This was a hard lesson
>> learned  (initially, by others, before my time) in the DoD in the early
>> 1990s, and that I personally experienced in the metadata wars of the
>> 2000s, where people will fight to the death to include their "words",
>> mistaking these for the concepts behind them.*
>
> This was a fight that the Cyc project thought it had solved in the
> early '90s by establishing the 2-way mappings between NL & ontology
> terms.  However, as Amanda well remembers, as new people came into
> Cycorp, there were continual fights over what a term in the ontology
> means -- and what it *should* mean, even when each of the competing
> meanings were worthwhile concepts to have in the ontology.  This has
> obviously continued after we both left.  You can find lots of terms in
> OpenCyc whose names could be interpreted in multiple ways, yet whose
> #$comments do little more than restate the name.  Often, the statements
> remaining about the concepts in OpenCyc are not sufficient to understand
> the intended meaning of the term.
>
>
>> The important point is
>> that if you shoehorn peoples'  semantics into "common" vocabularies
>> or ontologies (and it's worse  for the  latter, because then you
>> obliterate their differentiating semantics, not just how they refer to
>> things), they may say they will agree with that concensus, but they will
>> drag their feet, and the "information-sharing"  effort will fail.
>
> Many ontologies (e.g., the OBO ontologies) get around this by using
> numeric strings as IDs of the ontology terms, forcing people to look
> at various comments and alternate names to understand what is
> intended.
>
> Also, +1 to Amanda's points, below.
>
> -- doug foxvog
>
>
>> *This also means that you need to differentiate between stakeholder
>> committees and dedicated small ontology teams, where the latter work in
>> the background to define the  semantics they hear the committee members
>> (often domain experts) describe in the foreground.
>>
>> Thanks,
>> Leo
>>
>> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>> [mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Amanda
>> Vizedom
>> Sent: Thursday, November 22, 2012 1:23 PM
>> To: [ontolog-forum]
>> Subject: Re: [ontolog-forum] doing standards [was - Re: Webby objects]
>>
>> <climbs atop a small pile of soapboxes>
>>
>> (1) Controlling language -- setting and using naming standards, for
>> example, -- is *much* harder than building and using shared ontologies.
>> Why? Because language and naming are things humans do/use constantly,
>> are
>> part of their natural behavior, are intertwined with cognitive and
>> behavioral adaptations in many ways. And standardizing the language
>> people
>> use away from their ever-adapting expert language interferes with that
>> adaptation. If the job needs to be done, the tension here is (correctly)
>> going to resolve in favor of the adaptation and away from the
>> standardization.
>>
>> (2) It is a central, advantageous feature of ontological representation
>> that it can enable and support shared models, shared metadata, and
>> interoperability *without* requiring controlled and standardized
>> language
>> use. Not all ontology-based projects have staff with sufficient training
>> to understand and use that advantage. Those that don't use it end up
>> bogged down in naming and language debates.
>>
>> (3) Available out-of-the-box tools (and standard KR languages) need to
>> grow to support and emphasize this feature and ontologies (for example,
>> via built-in support for specifying multiple labels for a concept, with
>> lexical and localization information for those labels, and support for
>> viewing/browsing an ontology with some specific localization used to
>> select display labels. This is pretty easy compared to many other
>> ontology-interaction problems. But it isn't there in the off-the-shelf
>> stuff (in contrast, it *is* there in a range of commercial/proprietary
>> and/or in-house tools that have to deal with multilingual, business
>> vertical, or other localizations.
>>
>> (4) One reason people don't use ontologies more is because there is no
>> consensus on quality and methodology, and they don't know who to trust.
>> Those who say the have the One Methodology To Rule Them All are at best
>> no
>> better than everyone else at delivering something useful. That's because
>> we *don't* need One Methodology To Rule Them All. We need a good
>> understanding of what ontology features matter to what kinds of
>> projects:
>> that is, an understanding of what it means for an ontology to be
>> suitable
>> for an application, fit for a particular use case.
>>
>> (5) Until we build that kind of understanding -- at least a start --
>> some
>> people who could benefit from ontology applications will continue to
>> balk
>> at the lack of clarity, while other people will go ahead but invest in
>> ontology that is mismatched to the requirements. The latter experiences
>> make the climate for ontology adoption worse.
>>
>> (6) I suggest that it should be very high on our priority list to work
>> on
>> building that shared understanding of the relationships between use case
>> features and suitable ontology features. We can really only do that
>> effectively as a community, sharing experiences and collaborating on the
>> lessons learned. There are enough of us, across industries and sectors,
>> to
>> get going and make progress on this. There are also many who won't share
>> just yet because they aren't sure whether the payoff of sharing
>> outweighs
>> their trained resistance to making in-house knowledge public. We can
>> still
>> move, and try to pick them up as we gain momentum.
>>
>> (7) A good ontologist can, in fact, come in and build something and make
>> a
>> difference in two weeks. BUT this assumes that it is clear what the
>> ontology is for, how it will be used, what requirements come from
>> downstream, and what sources are to be regarded as authoritative, AND
>> that
>> the ontologists is provided access to those sources. The reality is that
>> the ontologist often walks into a situation in which few or none of
>> those
>> conditions are met. Sometimes it just hasn't occurred to anyone to do
>> them; sometimes it has but no one knows how. Sometimes an experienced
>> ontologist is actually there who knows how, but isn't give the authority
>> or resource access to do this requirements identification. Result: value
>> not delivered. I suggest that (6) is also important because it begins to
>> address this problem by building shared understanding of what use case
>> features are important to identify.
>>
>> </climbs down from soapboxes to go check on roasting turkey>
>>
>> Happy Thanksgiving!
>>
>> Best,
>> Amanda
>>
>>
>>
>> On Thu, Nov 22, 2012 at 12:17 PM, David Eddy
>
>> <deddy@xxxxxxxxxxxxx<mailto:deddy@xxxxxxxxxxxxx>> wrote:
>> Peter -
>>
>> On Nov 22, 2012, at 10:24 AM, Peter Yim wrote:
>>
>>
>> The basic question is "Why aren't practitioners using
>> ontologies?"
>>
>> Because they're really, really, really HARD, aside from the fact that
>> management has no idea what all the handwaving is about much less what
>> any
>> potential value might be.
>>
>> Plus the little detail of there being no foundation.  So ontologies are
>> very much ivory towers in the clouds, particularly in commercial IT.
>> Science/bio/pharma may very well be a different story since Mother
>> nature
>> has been carving the hierarchy for a few million years.  In commercial
>> systems the concept of ordered hierarchies & languages is at best a
>> distant fantasy.
>>
>>
>> It has been my direct, commercial experience with Fortune 500
>> organizations that very few (approaching none) have anything approaching
>
>> something simple as "naming standards" (conventions 3 ring
>> binders on
>
>> the shelf, yes.  Enforced no.).  Obviously some systems have naming
>> conventions, but that was typically driven by a single motivated
>> individual & is not carried over to the next systems(s).
>>
>> Please to acknowledge the scale here... a Fortune 500 scale organization
>> will have something like 1,000 to 5,000+ IBM mainframe applications.
>> This
>> is not counting client/server, web, *nix, iSeries, etc.  Pretty much
>> every
>> one of those applications will have uniquely opaque language.
>>
>>
>> I would argue that if the organization does not have the discipline to
>> control its language something as pie-in-the-sky as ontologies is a
>> castle
>> in the sand.
>>
>> The reasoning/experience is that if they're as big & important as they
>> are
>> without naming standards, why bother.
>>
>> It is my assumption the same thinking & non-action carries over to the
>> much more complex ontology domain.
>>
>> If the ontology consultant (that's singular) could show up on Monday &
>> create something substantive in a week or two, AND get easy-to-use &
>> understand results into the hands of the grunts, things would likely be
>
>> different.  Requiring the grunts the data
>> entry/programmer/analyst folks
>> doing the work to learn yet another collection of technical
>> languages is
>
>> not a route to adoption.
>>
>> Does anything in the ontology domain deliver something that's as
>> easy-to-use & mindlessly useful as a spellchecker?  Protege may be a
>> wonderful product, but it's not even remotely close to being an end-user
>> tool.
>>
>> - David
>>
>>
>>
>> _________________________________________________________________
>> 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<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
>>
>>
>>
>> _________________________________________________________________
>> 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
>>
>
>
>
> _________________________________________________________________
> 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
>
>
>
>
>
>
>
> --
> Edward J. Barkmeyer                        Email: edbark@xxxxxxxx
> National Institute of Standards & Technology
> Systems Integration Division, Engineering Laboratory
> 100 Bureau Drive, Stop 8263                Tel: +1 301-975-3528
> Gaithersburg, MD 20899-8263                Cel: +1 240-672-5800
>
> "The opinions expressed above do not reflect consensus of NIST,
>  and have not been reviewed by any Government authority."
>
> _________________________________________________________________
> 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
>



_________________________________________________________________
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

_________________________________________________________________
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
 

_________________________________________________________________
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    (01)