Re: How to force two entities to point to the same lookup value

thomas.kyte_at_oracle.com wrote:

> Bob Badour wrote:
>

>>>>>>>>>>It goes without saying that Tom Kyte says everyone should lock
>>>>>>>>>>themselves permanently into Oracle solutions. Oracle pays him to say
>>>>>>>>>>that,
>>>>>>>>>
>>>>>>>>I sincerely doubt that I would. I am sure he is a very pleasant man.
>>>>>>>>Regardless of his natural inclination, which I am sure is equally
>>>>>>>>pleasant, Oracle pays him to be pleasant as well.
>>>>>>>>

>>>>>>>[someone else said]
>>>>>>>The thing about Tom is, he goes through the technicalities
>>>>>>>point-by-point.

>>>>>>[bob said]
>>>>>>Does he? What are the points he uses to demonstrate that folks should
>>>>>>give his employer a monopolistic franchise over their business?
>>>>>

>>>>What a kooky thing to say! I mention the incentives an Oracle VP has for
>>>>getting people to give his company a monopolistic franchise over their
>>>>business, and you morph it into a technical discussion?
>>>

>>Are you suggesting there is anything illegal or morally wrong with
>>puffery? Are you suggesting that an appeal to authority is somehow
>>strong or stronger when one appeals to the authority of a shill?

I am very well-read on that subject, Tom. I am also very well-read on the logical fallacy of http://en.wikipedia.org/wiki/Appeal_to_authority

Are you familiar with Date's _Principle of Incoherence_ ? ie: "It is very difficult to respond coherently to that which is incoherent."

Relevant to this current discussion, it is hard to respond to the logical fallacy of "appeal to authority" without eliciting an accusation of "ad hominem". It hardly seems reasonable or valid to blame me for the quality of the material available for my reply.

> If you had ever bothered to read my books or whatever, you would
> understand that what I say is "hey, you paid a lot for your software -
> be it sqlserver, db2, informix, or even oracle. It is foolish not to
> use the tools you pay for".

Please bear with the length of what follows. The length is an exemplar of Date's _Principle of Incoherence_ cited above.

If the fee recurs (including forced upgrade fees), if purchase decisions are dependent on each other, if Oracle charges for expanded use, OR if use of a feature creates any expensive dependencies, what Tom Kyte says has no merit.

First, I wonder what sort of arrogance makes you suggest your own books are worth reading.

Next, I note your position statement here is brief enough Morgan could have easily substituted it for his appeal to authority, and anyone could have offered it as a reply to my direct question: "What are the points [Tom Kyte] uses to demonstrate that folks should give his employer a monopolistic franchise over their business?" Thank you for finally offering something from the oracle group that is neither logical fallacy nor psychotic non sequitur--even if long overdue and incomplete.

Finally, your position sounds good when you say it fast, but there are a number of factors that can and do render it completely invalid (or at least incompletely valid.) With respect to the issues of the logical fallacy of appeal to authority and of the motives of shills, I note in passing that companies pay shills both to say things that work in the company's interest and to not say things that work against the company's interest. The complete omission of any mention of the factors that can invalidate your position, which I explain below, renders what you wrote meaningless puffery. Your stated position is neither very true nor very false. Caveat emptor.

Your position has merit when purchase decisions are independent one-time decisions and when decisions to use features are independent or offer savings exceeding the cost of dependence. When "the tools you pay for" involve a recurring cost, the purchase decision is not a one-time decision. When "the tools you pay for" involve additional payments for expanded use, purchase decisions are not independent--especially when the vendor charges different customers different prices for the identical product. When the use of a product feature requires or prohibits the use of another product feature, feature use decisions are not independent.

As an example of the last point, I cannot use my vehicle's maximum acceleration and its maximum fuel economy at the same time. However, the cost of switching between those decisions is negligible so I can use both features at different times independently. My vehicle's durability and beauty, on the other hand, are not as independent. If I make maximum use of my vehicle's impact resistance, it will then cost a great deal to make maximum use of my vehicle's aesthetic charm. When I paid for the vehicle, I got both.

As an example of the third point, my vehicle requires no additional payments for expanded use (ignoring fuel and wear for the moment). I can use it a little or a lot. I can drive it to another province or another country or I can lend it to a friend, and neither the company I bought it from nor the company that manufactured it charges me any extra. Contrast this with hiring an electrical contractor when building a building. The contractor specifies a price for what's on the drawings. If I later decide I want the contractor to install more light fixtures near the entrances and the parking lots, I have to pay extra for the expanded use of the contractor's service. Having once written estimating software for electrical contractors, I note that competitive bids are bid very near (sometimes slightly below) cost and change notices are bid at significant multiples to cost.

To illustrate the dependence of purchase decisions, I note that my vehicle purchase decisions are independent of each other. Having a Ford today and a Nissan yesterday and a ??? tomorrow don't introduce any costs. My vehicle purchases do, however, affect my insurance purchases and can affect my fuel purchases and vice versa. My choice of mechanic can affect my vehicle purchases and vice versa. My choice of cell-phone can affect my choice of carrier and vice versa. For the most part, however, choosing to use a feature of any of those things won't affect my costs. Suppose, for example, I choose to use the small size of my compact car when building a garage. That decision will affect my future purchase decisions or my future use of the garage. Suppose, for example, I choose to use the load capacity of my 1-ton truck when installing a wood-burning furnace. Those decisions create dependencies that can involve significant future costs.

To illustrate the impact of recurring fees, consider the land for a building site. Suppose the land offers existing AC electric service, reasonably level ground and good solid clay. I can use the good solid clay to dig a basement, pour a foundation and build a rock-solid house. Or I can use the reasonably level ground to park a mobile-home on the land. In either case, I can use the existing AC service to power my appliances. Either way, I am paying for the ground so I might make maximum use of it and build a rock-solid house. Right? But there is a big difference between buying the ground and leasing the ground. If I am leasing the ground, I have a recurring cost. If I dig a basement and pour a foundation, I must either continue to pay that recurring cost indefinitely or I must pay a large expense to move the house to a new site abandoning the foundation or I must abandon the house. If I put a mobile home on the site, I am deciding not to use the good solid building clay, but I can easily move the home to a new site at the end of any lease period. My decision to use the alternating current service has fewer repercussions because many other building sites offer existing alternating current service or new service at reasonable cost, and if necessary, I can switch to DC simply by buying a suitable inverter.

To decide whether what Tom Kyte says has merit, one must ask:

• Is the Oracle license a one-time purchase or a recurring subscription fee? Even if nominally one-time, will anything force recurring upgrade fees?
• Are my dbms purchase decisions independent of each other? If I decide to create another system using a dbms, will I want to use what I already have?
• Does Oracle charge an additional fee for expanded use like additional users or additional servers?
• Does the use decision force any dependencies? If so, at what cost? Does using a feature limit or preclude the use of any other features? If so, at what cost?

If the fee recurs (including forced upgrade fees), if purchase decisions are dependent on each other, if Oracle charges for expanded use, OR if use of a feature creates any expensive dependencies, what Tom Kyte says has no merit.

SDO forces dependencies on Oracles PL/SQL stored procedures and on Oracle's proprietary design. If SDO is exactly what someone was going to create on their own and if they can use it without any further effort or cost, then the decision to use it is obvious. Even if one must later pay the price of re-creating to use a different dbms, one has at least deferred that cost. However, if SDO is not exactly what one was going to create, then whether to use it depends on 1) the incremental cost of using it, 2) the incremental cost of switching to a different vendor if necessary and 3) the probability of needing to switch.

Given that the original poster expressed concern about cross-vendor portability, I assume he rationally weighed the factors I mention (and possibly others) to decide that the potential to switch vendors has some attraction. From what I read about SDO, it bears no resemblance to the original poster's requirements suggesting significant cost to use and significant cost to refit to a different vendor.

   (i actually wrote words exactly to that
> effect, yes, an Oracle employee wrote "use the heck out of the software
> you buy, regardless of whom you bought it from")

You are being disingenuous. Drug dealers emphasize the cost of free samples to their potential customers and the availability of liberal credit to their potential distributors too--regardless of who is offering the samples or the credit.

> Do I teach people how to maximize their investment in the product
> called Oracle - absolutely, that is part of my job (some might even say
> my "hobby"). Sqlserver technical folks do it for sqlserver, IBM folks
> for DB2 and Informix.

Are you suggesting that the existence of shills for other companies makes you any less a shill or makes appeals to their authority any less fallacious?

   Some people actually desire this, find this
> desirable (maximizing their return on investment, making the most of
> the products they purchase).

If what you wrote had had merit instead of being mere puffery, believing that your writing maximizes customer ROI would be rational. As it is, doing so is not rational. Lucky for you, not everyone is rational.

> Do I also work in the software/IT industry in general? Absolutely.
> Implementing software solutions is part of my job.

I am sure that gives you plenty of opportunity to implement solutions that lock people into your company's software. I doubt Oracle would have made you a VP if you locked people out instead.

> If you have to say things like "it goes without saying", you have a
> weak argument.

If you have to say, "Tom Kyte said so and if you read every word he ever wrote you would agree" is much, much weaker. Given that what was suggested was predictably exactly what any other Oracle shill would write, we could replace it with "An Oracle shill said so and if you read every word he ever wrote you would agree" without changing the meaning or import of the words. The logical fallacy of an appeal to authority is very weak. Combined with obvious motive for bias an appeal to authority is even weaker.

   You are stating something as fact without any
> supporting information whatsoever.

The supporting information was the "VP Oracle" after your name and the observable facts regarding who would likely benefits from the suggestion vs. who might be burdened by it. I don't recall whether the "VP Oracle" was presented in this thread, but it is readily available to anyone who looks. It's probably the first thing anyone would see if they were trying to "read what Tom Kyte wrote" as suggested to me (your publishers seem to think it enhances your credibility or something.)

The observable facts regarding costs and benefits were 1) DA Morgan's post, 2) the documents at the other end of the link he gave, and 3) the original poster's original post. It doesn't take much to infer the balance of benefits favours Oracle and to infer the balance of costs burdens the original poster. It really doesn't.

   I've seen some of your other
> arguments regarding database design and they in general are strong -
> you should know better.

I will leave it for others to decide who should know better.

> When something "goes without saying" - it really doesn't. You had to
> say it - it couldn't go without saying (sort of like saying "it is
> obvious that...", if it were obvious, you wouldn't need to state it)

But I am not the one who said it: DA Morgan did. I merely pointed out how superfluous (and hence pointless) it was when he said it.

> I agree that "appeal to authority" does not a case make, but you stand
> on the same ground with the ad hominem "shill" and "it goes without
> saying".

My case was never limited to that. My case always included the observable fact that the suggestion was very much in Oracle's financial interest and very likely contrary to the original poster's interest.

The fact that you are a shill and that shills advance their companys' interests without quite as much regard for their prospects interests sufficed to demonstrate that I probably would not have to read everything you ever wrote to understand (and dismiss) your position--despite DA Morgan's suggestion to that effect. To use it to demonstrate risk of bias is not a fallacy at all.

Combining apparent bias with a logical fallacy sets the bar pretty low. Wouldn't you agree?

> I presume you are a software developer of some sort - or at least
> involved in the industry. When someone uses a
> tool/application/whatever designed and developed by you or your company
> do you:

Do I shill my own products? Yes. I exhibit bias regarding my products and services. I think everyone should have all of them, and those who have all of them should have multipe copies of them. If you have never bought any, you should. If you have, buy more.

> a) show them how to best take advantage of the features so as to make
> it perform, scale, function well?

Yes.

> b) tell them not to use 90% of the functionality since they won't be
> able to switch to another tool/application? (do no evil right).

Non sequitur. Unless you are saying that SDO is 90% of Oracle's functionality, which I think many of your customer's would find surprising.

Unlike many software developers, I have always made sure my services and products are as replaceable as possible. (I suck at marketing.)

> A) - that is what I do with Oracle. I show people how to maximize the
> product they just purchased.

I disagree. If your focus were not on maximizing value for Oracle's shareholders, you would not be a VP. Suggesting that you do anything else is disingenuous.

Insofar as Oracle's customers' interests sometimes align with Oracle's shareholders', what you do will sometimes maximize their value. Insofar as Oracle's customers' interests sometimes conflict with Oracle's shareholders', what you do will sometimes cost them.

There is plenty of empirical evidence here that encouraging the original poster to build a solution around SDO may cost that customer more than it benefits him. It doesn't take long for an objective third party, such as myself, to reach that conclusion. It takes a lot more than "read every book written by some Oracle VP" to make me conclude otherwise.

   I make the more general point as well
> that you *paid for this stuff, regardless of who you paid it to*, use
> it, don't let it just sit there, use it. You seem to think that evil,
> so be it.

Now who is using ad hominem? The merit of your position, as established at length above, is tenuous at best given the factors you omit. I base my position on costs and benefits not good and evil, which means even your fallacious argument lacks a key component.

[puffery snipped]

> I have a feeling you'll never be convinced that I'm not a 'shill', I
> don't really care one way or the other.

Strictly speaking, 'shill' is poor diction. You divulge your employment by Oracle, so you are technically not a shill. Insofar as you pretend to advance any interest other than your own or the interest of Oracle's shareholders, you are not too different from a shill though. While a more accurate term would be 'hack', I exercise a little license for the sake of communication. Readers more reliably understand that the issue is hidden or denied bias when they read 'shill'. Some would assume 'hack' refers to the quality of your writing rather than the motivation for it.

   Just felt the need to point
> out the logical fallacy in your argument here. You've done it well
> yourself in the past.

No problem. Indulge your feelings all you want. I don't mind. It's not like the logical fallacy was at the core of my argument or anything.

[more puffery snipped]

> Thomas Kyte
> http://asktom.oracle.com/
> Resistance is Futile (that of course is an attempt at humor, a bit of
> which this thread is in dire need of)...

Hey! I resemble that remark! I happen to find frangible woozle muffs very humorous, and I find a lot of what is written in this thread funny as all hell. The appearance that some of the humour might be unintended only enhances it.

P.S. Now that you have had the opportunity to come forward and present your position, I have to say the empirical evidence only confirms that bias does render what you write incoherent. Since life is short and since I don't have time to compose coherent replies to every effortless incorehent post, I am adding you to my killfile. Plonk! Received on Sat Aug 19 2006 - 01:20:10 CEST