Re: Cost/Benefit Anyone? Re: Vote for my Semantic Web presentation at SXSW

As fascinating as this discussion is, maybe the two of you want to
work it out directly and then report back with a summary?

Speaking as just one subscriber's data point, of course, I'm...

-Patrick


On Fri, Aug 19, 2011 at 7:41 AM, Patrick Durusau <patrick@durusau.net> wrote:
> Kingsley,
>
> Correction: I have never accused you of being modest or of not being an
> accountant. ;-)
>
> Nor have I said the costs you talk about in your accountant voice don't
> exist.
>
> The problem is identifying the cost to a particular client, say of email
> spam, versus the cost the solution for the same person.
>
> For example, I picked a spam article at random that says a 100 person firm
> *could be losing* as much as $55,000 per year due to spam.
>
> Think about that for a minute. That works out to $550 per person.
>
> So, if your solution costs more than $550 per person, it isn't worth buying.
>
> Besides, the $550 per person *isn't on the books.* Purchasing your solution
> is. As they say, spam is a hidden cost. Hidden costs are hard to quantify or
> get people to address.
>
> Not to mention that your solution requires an investment before the software
> can exist for any benefit. That is an even harder sell.
>
> Isn't investment to enable a return from another investment (software,
> later) something accountants can see?
>
> Hope you are having a great day!
>
> Patrick
>
>
> PS: The random spam article:
> http://blogs.cisco.com/smallbusiness/the_big_cost_of_spam_viruses_for_small_business/
>
>
> On 8/19/2011 9:57 AM, Kingsley Idehen wrote:
>>
>> On 8/19/11 6:37 AM, Patrick Durusau wrote:
>>>
>>> Kingsley,
>>>
>>> One more attempt.
>>>
>>> The "press release" I pointed to was an example that would have to be
>>> particularized to a CIO or CTO in term of *their* expenses of integration,
>>> then showing *their* savings.
>>
>> Yes, and I sent you a link to a collection of similar documents from which
>> you could find similar research depending on problem type. On the first page
>> you should have seen a link to a research document about the cost of email
>> spam, for instance.
>>
>> CEO, CIOs, CTOs are all dealing with costs of:
>>
>> 1. Spam
>> 2. Password Management
>> 3. Security
>> 4. Data Integration.
>>
>> There isn't a shortage of market research material re. the above and their
>> costs across a plethora of domains.
>>
>>>
>>> The difference in our positions, from my "context," is that I am saying
>>> the benefit to enterprises has to be expressed in terms of *their* bottom
>>> line, over the next quarter, six months, year.
>>
>> For what its worth I worked for many years as an accountant before I
>> crossed over to the vendor realm during the early days of Open Systems --
>> when Unix was being introduced to enterprises. That's the reason why
>> integration middleware and dbms technology has been my passion for 20+
>> years. I am a slightly different profile to what you assume in your comments
>> re. cost-benefits analysis.
>>
>>> I "hear" (your opinion likely differs) you saying there is a global
>>> benefit that enterprises should invest in with no specific ROI for their
>>> bottom line in any definite period.
>>
>> See comment above. I live problems first, then architect technology to
>> solve them. When I tell you about the costs of data integration to
>> enterprises I am basically telling you that I've lived the problem for many
>> years. My understanding is quite deep. Sorry, but this isn't an area when I
>> can pretend to be modest :-)
>>
>>>
>>> Case in point, CAS, http://www.cas.org/. Coming up on 62 million organic
>>> and inorganic substances given unique identifiers. What is the incentive for
>>> any of their users/customers to switch to Linked Data?
>>
>> I think the issue is more about: what would identifiers provide to this
>> organization with regards to the obvious need to virtualize its critical
>> data sources such that:
>>
>> 1. data sources are represented as fine grained data objects
>> 2. every data object is endowed with an identifier
>> 3. identifiers become superkey that provide conduits highly navigable data
>> object based zeitgeists -- a single identifier should resolve to graph
>> pictorial representing all data associated with that specific identifier and
>> and additional data that has been reconciled logically e.g., leverage
>> owl:sameAs and IFP (inverse functional property) logic.
>>
>>>
>>> As I said several post ago, your success depends upon people investing in
>>> a technology for your benefit. (In all fairness you argue they benefit as
>>> well, but they are the best judges of the best use of their time and
>>> resources.)
>>
>> Kingsley
>>>
>>> Hope you are looking forward to a great weekend!
>>>
>>> Patrick
>>>
>>> On 8/18/2011 10:09 PM, Kingsley Idehen wrote:
>>>>
>>>> On 8/18/11 5:27 PM, Patrick Durusau wrote:
>>>>>
>>>>> Kingsley,
>>>>>
>>>>> Citing your own bookmark file hardly qualifies as market numbers.
>>>>
>>>> My own bookmark? I gave you a URL to a bookmark collection. The
>>>> collection contains links for a variety of research documents.
>>>>
>>>>> People promoting technologies make up all sorts of numbers about what
>>>>> use of X will save. Reminds me of the music or software theft numbers.
>>>>
>>>> Er. and you posted a link to a press release. What's your point?
>>>>
>>>>> They have no relationship to any reality that I share.
>>>>
>>>> But you posted an Informatica press release to make some kind of point.
>>>> Or am I completely misreading and misunderstanding the purpose of that URL
>>>> too?
>>>>
>>>>>
>>>>> It's been enjoyable as usual but without some common basis for
>>>>> discussion we aren't going to get any closer to a common understanding.
>>>>
>>>> Correct :-)
>>>>
>>>> Kingsley
>>>>>
>>>>> Hope you are having a great week!
>>>>>
>>>>> Patrick
>>>>>
>>>>>
>>>>>
>>>>> On 8/18/2011 3:24 PM, Kingsley Idehen wrote:
>>>>>>
>>>>>> On 8/18/11 2:50 PM, Patrick Durusau wrote:
>>>>>>>
>>>>>>> Kingsley,
>>>>>>>
>>>>>>> On 8/18/2011 1:52 PM, Kingsley Idehen wrote:
>>>>>>>>
>>>>>>>> On 8/18/11 1:40 PM, Patrick Durusau wrote:
>>>>>>>>>
>>>>>>>>> Kingsley,
>>>>>>>>>
>>>>>>>>> From below:
>>>>>>>>>
>>>>>>>>>> This critical value only materializes via appropriate "context
>>>>>>>>>> lenses". For decision makers it is always via opportunity costs.  If someone
>>>>>>>>>> else is eating you lunch by disrupting your market you simply have to
>>>>>>>>>> respond. Thus, on this side of the fence its better to focus on eating lunch
>>>>>>>>>> rather than warning about the possibility of doing so, or outlining how it
>>>>>>>>>> could be done. Just do it!
>>>>>>>>>
>>>>>>>>> I appreciate the sentiment, "Just do it!" as my close friend Jack
>>>>>>>>> Park says it fairly often.
>>>>>>>>>
>>>>>>>>> But "Just do it!" doesn't answer the question of cost/benefit.
>>>>>>>>
>>>>>>>> I mean: just start eating the lunch i.e., make a solution that takes
>>>>>>>> advantage of an opportunity en route to market disruption. Trouble with the
>>>>>>>> Semantic Web is that people spend too much time arguing and postulating.
>>>>>>>> Ironically, when TimBL worked on the early WWW, his mindset was: just do it!
>>>>>>>> :-)
>>>>>>>>
>>>>>>> Still dodging the question I see. ;-)
>>>>>>
>>>>>> Of course not.
>>>>>>
>>>>>> You want market research numbers, see the related section at the end
>>>>>> of this reply. I sorta assumed you would have found this serendipitously
>>>>>> though? Ah! You don't quite believe in the utility of this Linked Data stuff
>>>>>> etc..
>>>>>>
>>>>>>>
>>>>>>>>>
>>>>>>>>> It avoids it in favor of advocacy.
>>>>>>>>
>>>>>>>> See my comments above. You are skewing my comments to match you
>>>>>>>> desired outcome, methinks.
>>>>>>>>
>>>>>>> You reach that conclusion pretty frequently.
>>>>>>
>>>>>> See my earlier comment.
>>>>>>
>>>>>>>
>>>>>>> I ask for hard numbers, you say that isn't your question and/or
>>>>>>> skewing your comments.
>>>>>>
>>>>>> Yes. I didn't know this was about market research and numbers [1].
>>>>>>
>>>>>>>
>>>>>>>>>
>>>>>>>>> Example: Privacy controls and Facebook. How much would it cost to
>>>>>>>>> solve this problem?
>>>>>>>>
>>>>>>>> I assume you know the costs of the above.
>>>>>>>> It won't cost north of a billion dollars to make a WebID based
>>>>>>>> solution. In short, such a thing has existed for a long time, depending on
>>>>>>>> your "context lenses" .
>>>>>>>>
>>>>>>>
>>>>>>> I assume everyone here is familiar with: http://www.w3.org/wiki/WebID
>>>>>>> ?
>>>>>>>
>>>>>>> So we need to take the number of users who have a WebID and subtract
>>>>>>> that from the number of FaceBook users.
>>>>>>>
>>>>>>> Yes?
>>>>>>
>>>>>> No!
>>>>>>
>>>>>> Take the number of people that have are members of a service that's
>>>>>> ambivalent to the self calibration of the vulnerabilities of its members
>>>>>> aka. privacy.
>>>>>>
>>>>>>>
>>>>>>> The remaining number need a WebID or some substantial portion, yes?
>>>>>>
>>>>>> Ultimately they need a WebID absolutely! And do you know why? It will
>>>>>> enable members begin the inevitable journey towards self calibration of
>>>>>> their respective vulnerabilities.
>>>>>>
>>>>>> I hope you understand that society is old and the likes of G+, FB are
>>>>>> new and utterly immature. In society, one is innocent until proven guilty or
>>>>>> not guilty. In the world of FB and G+ the fundamentals of society are
>>>>>> currently being inverted. Anyone can ultimately say anything about you. Both
>>>>>> parties are building cyber police states via their respective silos. Grr...
>>>>>> don't get me going on this matter.
>>>>>>
>>>>>> Every single netizen needs a verifiable identifier. That's the bottom
>>>>>> line, and WebID (courtesy of Linked Data) and Trust Semantics nails the
>>>>>> issue.
>>>>>>
>>>>>>>
>>>>>>> So who bears that cost? Each of those users? It cost each of them
>>>>>>> something to get a WebID. Yes?
>>>>>>
>>>>>> Look here is a real world example. Just google up on wire shark re.
>>>>>> Facebook and Google. Until the wire shark episodes both peddled lame excuses
>>>>>> for not using HTTPS. Today both use HTTPS. Do you want to know why? Simple
>>>>>> answer: opportunity cost of not doing so became palpable.
>>>>>>
>>>>>>>
>>>>>>> What is their benefit from getting that WebID? Will it outweigh their
>>>>>>> cost in their eyes?
>>>>>>
>>>>>> See comment above.
>>>>>>
>>>>>> We've already witnessed Craigslist horrors. But all of this is child's
>>>>>> play if identity isn't fixed on the InterWeb. If you think I need to give
>>>>>> you market numbers for that too, then I think we are simply talking past
>>>>>> ourselves (a common occurence).
>>>>>>
>>>>>>>
>>>>>>>>> Then, what increase in revenue will result from solving it?
>>>>>>>>
>>>>>>>> FB -- less vulnerability and bleed.
>>>>>>>>
>>>>>>>> Startups or Smartups: massive opportunity to make sales by solving a
>>>>>>>> palpable problem.
>>>>>>>>
>>>>>>>>>
>>>>>>>>> Or if Facebook's lunch is going to be eaten, say by G+, then why
>>>>>>>>> doesn't G+ solve the problem?
>>>>>>>>
>>>>>>>> G+ is trying to do just that, but in the wrong Web dimension. That's
>>>>>>>> why neither G+ nor FB have been able to solve the identity reconciliation
>>>>>>>> riddle.
>>>>>>>>
>>>>>>> Maybe you share your observations with G and FB. ;-)
>>>>>>
>>>>>> Hmm. wondering how you've concluded either way :-)
>>>>>>
>>>>>>>
>>>>>>> Seriously, I don't think they are as dumb as everyone seems to think.
>>>>>>
>>>>>> I haven't characterized them as dumb. I would put this in the careless
>>>>>> and ambivalent bucket.
>>>>>>
>>>>>>>
>>>>>>> It may well be they have had this very discussion and decided it
>>>>>>> isn't cost effective to address.
>>>>>>
>>>>>> See my earlier comments. Or just look at the G+ "real names"
>>>>>> imbroglio.
>>>>>>
>>>>>>>
>>>>>>>>>
>>>>>>>>> Are privacy controls are a non-problem?
>>>>>>>>>
>>>>>>>>> Your "context lenses."
>>>>>>>>>
>>>>>>>>> True, you can market a product/service that no one has ever seen
>>>>>>>>> before. Like pet rocks.
>>>>>>>>>
>>>>>>>>> And they "just did it!"
>>>>>>>>>
>>>>>>>>> With one important difference.
>>>>>>>>>
>>>>>>>>> Their *doing it* did not depend upon the gratuitous efforts of
>>>>>>>>> thousands if not millions of others.
>>>>>>>>
>>>>>>>> Don't quite get your point. I am talking about a solution that
>>>>>>>> starts off with identity reconciliation, passes through access control
>>>>>>>> lists, and ultimately makes virtues of heterogeneous data virtualization
>>>>>>>> clearer re. data integration pain alleviation.
>>>>>>>>
>>>>>>>> In the above we have a market place north of 100 Billion Dollars.
>>>>>>>>
>>>>>>>
>>>>>>> Yes, but your solution: "...starts off with identity
>>>>>>> reconciliation..."
>>>>>>
>>>>>> See comments above about WebID and Trust Logic. It just another way of
>>>>>> referring to the issues that have resulted in outputs from the Semantic Web
>>>>>> Project.
>>>>>>
>>>>>>>
>>>>>>> Sure, start with the critical problem already solved and you really
>>>>>>> are at a "...market place north of 100 Billion Dollars...", but that is all
>>>>>>> in your imagination.
>>>>>>
>>>>>> See my earlier comments. And for your numbers, see links below.
>>>>>>>
>>>>>>> Having a system of assigned and reconciled WebIDs isn't a zero cost
>>>>>>> to users or businesses solution. It is going to cost someone to assign and
>>>>>>> reconcile those WebIDs. Yes?
>>>>>>
>>>>>> You can buy a solution that post installation will make and reconcile
>>>>>> all kinds of identifiers including those that serve are WebIDs for humans or
>>>>>> agents.
>>>>>>
>>>>>>>
>>>>>>> Since it is your solution, may I ask who is going to pay that cost?
>>>>>>
>>>>>> Companies have been paying for it already, for quite some time :-) I
>>>>>> am not speculating, simply sharing perspective re. what commercial
>>>>>> opportunities exist when you grok the Semantic Web Project stack and the
>>>>>> application of its output to solutions that solve real problems.
>>>>>>
>>>>>>
>>>>>>>
>>>>>>>>>
>>>>>>>>> Isn't that an important distinction?
>>>>>>>>
>>>>>>>> Yes, and one that has never been lost on me :-)
>>>>>>>>
>>>>>>>
>>>>>>> Interested to hear your answer since that distinction has never been
>>>>>>> lost on you.
>>>>>>
>>>>>> Links:
>>>>>>
>>>>>> 1. http://www.delicious.com/kidehen/market_research -- I am sure you
>>>>>> can filter through
>>>>>>
>>>>>>
>>>>>> Kingsley
>>>>>>>
>>>>>>> Patrick
>>>>>>>
>>>>>>>
>>>>>>>> Kingsley
>>>>>>>>>
>>>>>>>>> Hope you are having a great day!
>>>>>>>>>
>>>>>>>>> Patrick
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On 8/18/2011 10:54 AM, Kingsley Idehen wrote:
>>>>>>>>>>
>>>>>>>>>> On 8/18/11 10:25 AM, Patrick Durusau wrote:
>>>>>>>>>>>
>>>>>>>>>>> Kingsley,
>>>>>>>>>>>
>>>>>>>>>>> Your characterization of "problems" is spot on:
>>>>>>>>>>>
>>>>>>>>>>> On 8/18/2011 9:01 AM, Kingsley Idehen wrote:
>>>>>>>>>>>
>>>>>>>>>>> <snip>
>>>>>>>>>>>
>>>>>>>>>>>> Linked Data addresses many real world problems. The trouble is
>>>>>>>>>>>> that problems are subjective. If you have experienced a problem it doesn't
>>>>>>>>>>>> exist. If you don't understand a problem it doesn't exist. If you don't know
>>>>>>>>>>>> a problem exists then again it doesn't exist in you context.
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> But you left out: The recognized "problem" must *cost more* than
>>>>>>>>>>> the cost of addressing it.
>>>>>>>>>>
>>>>>>>>>> Yes. Now in my case I assumed the above to be implicit when
>>>>>>>>>> context is about a solution or solutions :-)
>>>>>>>>>>
>>>>>>>>>> If a solution costs more than the problem, it is a problem^n
>>>>>>>>>> matter. No good.
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> A favorable cost/benefit ratio has to be recognized by the people
>>>>>>>>>>> being called upon to make the investment in solutions.
>>>>>>>>>>
>>>>>>>>>> Always! Investment evaluation 101 for any business oriented
>>>>>>>>>> decision maker.
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> That is recognition of a favorable cost/benefit ratio by the W3C
>>>>>>>>>>> and company is insufficient.
>>>>>>>>>>>
>>>>>>>>>>> Yes?
>>>>>>>>>>
>>>>>>>>>> Yes-ish. And here's why. Implementation cost is a tricky factor,
>>>>>>>>>> one typically glossed over in marketing communications that more often than
>>>>>>>>>> not blind side decision makers; especially those that are extremely
>>>>>>>>>> technically challenged. Note, when I say "technically challenged" I am not
>>>>>>>>>> referring to programming skills. I am referring to basic understanding of
>>>>>>>>>> technology as it applies to a given domain e.g. the enterprise.
>>>>>>>>>>
>>>>>>>>>> Back to the W3C and "The Semantic Web Project". In this case, the
>>>>>>>>>> big issue is that degree of unobtrusive delivery hasn't been a leading
>>>>>>>>>> factor -- bar SPARQL where its deliberate SQL proximity is all about
>>>>>>>>>> unobtrusive implementation and adoption. Ditto R2RML .
>>>>>>>>>>
>>>>>>>>>> RDF is an example of a poorly orchestrated revolution at the
>>>>>>>>>> syntax level that is implicitly obtrusive at adoption and implementation
>>>>>>>>>> time. It is in this context I agree fully with you. There was a
>>>>>>>>>> misconception that RDF would be adopted like HTML, just like that. As we can
>>>>>>>>>> all see today, that never happened and will never happened via revolution.
>>>>>>>>>>
>>>>>>>>>> What can happen, unobtrusively, is the use and appreciation of
>>>>>>>>>> solutions that generate Linked Data (expressed using a variety of syntaxes
>>>>>>>>>> and serialized in a variety of formats). That's why we've invested so much
>>>>>>>>>> time in both Linked Data Middleware and DBMS technology for ingestion,
>>>>>>>>>> indexing, querying, and serialization.
>>>>>>>>>>>
>>>>>>>>>>>> For the umpteenth time here are three real world problems
>>>>>>>>>>>> addressed effectively by Linked Data courtesy of AWWW (Architecture of the
>>>>>>>>>>>> World Wide Web):
>>>>>>>>>>>>
>>>>>>>>>>>> 1. Verifiable Identifiers -- as delivered via WebID (leveraging
>>>>>>>>>>>> Trust Logic and FOAF)
>>>>>>>>>>>> 2. Access Control Lists -- an application of WebID and Web
>>>>>>>>>>>> Access Control Ontology
>>>>>>>>>>>> 3. Heterogeneous Data Access and Integration -- basically taking
>>>>>>>>>>>> use beyond the limits of ODBC, JDBC etc..
>>>>>>>>>>>>
>>>>>>>>>>>> Let's apply the items above to some contemporary solutions that
>>>>>>>>>>>> illuminate the costs of not addressing the above:
>>>>>>>>>>>>
>>>>>>>>>>>> 1. G+ -- the "real name" debacle is WebID 101 re. pseudonyms,
>>>>>>>>>>>> synonyms, and anonymity
>>>>>>>>>>>> 2. Facebook -- all the privacy shortcomings boil down to not
>>>>>>>>>>>> understanding the power of InterWeb scale verifiable identifiers and access
>>>>>>>>>>>> control lists
>>>>>>>>>>>> 3. Twitter -- inability to turn Tweets into structured
>>>>>>>>>>>> annotations that are basically nano-memes
>>>>>>>>>>>> 4. Email, Comment, Pingback SPAM -- a result of not being able
>>>>>>>>>>>> to verify identifiers
>>>>>>>>>>>> 5. Precision Find -- going beyond the imprecision of Search
>>>>>>>>>>>> Engines whereby subject attribute and properties are used to contextually
>>>>>>>>>>>> discover relevant things (explicitly or serendipitously).
>>>>>>>>>>>>
>>>>>>>>>>>> The problem isn't really a shortage of solutions, far from it.
>>>>>>>>>>>>
>>>>>>>>>>> For the sake of argument only, conceding these are viable
>>>>>>>>>>> solutions, the question is:
>>>>>>>>>>>
>>>>>>>>>>> Do they provide more benefit than they cost?
>>>>>>>>>>
>>>>>>>>>> Yes. They do, unequivocally.
>>>>>>>>>>>
>>>>>>>>>>> If that can't be answered favorably, in hard currency (or some
>>>>>>>>>>> other continuum of value that appeals to particular investors), no one is
>>>>>>>>>>> going to make the investment.
>>>>>>>>>>>
>>>>>>>>>>> Economics 101.
>>>>>>>>>>
>>>>>>>>>> This critical value only materializes via appropriate "context
>>>>>>>>>> lenses". For decision makers it is always via opportunity costs.  If someone
>>>>>>>>>> else is eating you lunch by disrupting your market you simply have to
>>>>>>>>>> respond. Thus, on this side of the fence its better to focus on eating lunch
>>>>>>>>>> rather than warning about the possibility of doing so, or outlining how it
>>>>>>>>>> could be done. Just do it!
>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> That isn't specific to SemWeb but any solution to a problem.
>>>>>>>>>>
>>>>>>>>>> Yes!
>>>>>>>>>>
>>>>>>>>>>> The solution has to provide a favorable cost/benefit ratio or it
>>>>>>>>>>> won't be adopted. Or at least not widely.
>>>>>>>>>>>
>>>>>>>>>>> Hope you are having a great day!
>>>>>>>>>>>
>>>>>>>>>>> Patrick
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>>
>
> --
> Patrick Durusau
> patrick@durusau.net
> Chair, V1 - US TAG to JTC 1/SC 34
> Convener, JTC 1/SC 34/WG 3 (Topic Maps)
> Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300
> Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps)
>
> Another Word For It (blog): http://tm.durusau.net
> Homepage: http://www.durusau.net
> Twitter: patrickDurusau
>
>
>

Received on Friday, 19 August 2011 16:48:07 UTC