 | |
Re: some ideas about db rheory
comp.databases.theory comp.databases.theory
 |
| | Thread Tools | Display Modes |
#11
 
| |||
| |||
|
Re: some ideas about db rheory -
07-28-2009
, 12:24 AM
|
I wrote: [...] if you have no way to track identity across changes in real life, adding it as a modeling feature (either with explicit identities or by distinguishing between updates and deletes+inserts, as Brian proposes) isn't going to help a bit. Brian replies: [...] Either there can be change, which implies that there can be things that can change, or there cannot be change, which means that there cannot be deletes or inserts, let alone updates. No, Brian. These deletes, inserts, and updates are about statements of fact about the world, which can change to reflect new or changed observations, even when we haven't identified any objects that these statements are about to the extent you appear to deem necessary. I can observe Mary's goldfish and its medical condition, and truthfully record that in my database, two days in a row, *regardless of* whether I can tell whether we're dealing with the same goldfish in both cases. We may choose not to care and it won't be a problem. But for goldfish, at least we know they are identifiable in principle. This is not so easy for other types of objects; say, species of fish, countries of the world, or clouds. |
somehow,
It's cloud illusions I recall. I really don't know clouds, at all.
#12
| |||||
| |||||
|
|
I wrote: [...] if you have no way to track identity across changes in real life, adding it as a modeling feature (either with explicit identities or by distinguishing between updates and deletes+inserts, as Brian proposes) isn't going to help a bit. Brian replies: [...] * Either there can be change, which implies that there can be things that can change, or there cannot be change, which means that there cannot be deletes or inserts, let alone updates. No, Brian. *These deletes, inserts, and updates are about statements of fact about the world, which can change to reflect new or changed observations, even when we haven't identified any objects that these statements are about to the extent you appear to deem necessary. *I can observe Mary's goldfish and its medical condition, and truthfully record that in my database, two days in a row, *regardless of* whether I can tell whether we're dealing with the same goldfish in both cases. |
that can change, or there cannot be change, which means that there
cannot be deletes or inserts, let alone updates. Deletes, inserts and
updates are not about statements, they are themselves statements that
assert what is different and how it is different; therefore, if there
cannot be change, then nothing can be different and thus there cannot
be any deletes, inserts or updates. The implication is that if there
cannot be change, then there can be absolutely no relationship between
databases at different times, so the concept of deleting would not
make sense, for it assumes a relationship between databases, neither
would that of inserting, and most definitely not updating.
Your argument neither affirms or denies my statement. What it does is
illustrate the fact that descriptions, such as "Mary's goldfish," can
refer to different things at different times in the same way that a
particular instance of a composite key can refer to different things
at different times--something I've been arguing for a long time.
Because interpreters are temporal beings, interpretations must occur /
at a time/, which requires that there must be different
interpretations at different times. Even under the possible worlds
paradigm, only one possible world at a time can be the actual world.
Quote:
|
We may choose not to care and it won't be a problem. But for goldfish, at least we know they are identifiable in principle. This is not so easy for other types of objects; say, species of fish, countries of the world, or clouds. Once you commit to change, a system that has no way to track identity across changes is broken. If you mean: once you commit to being able to identify objects across changes, you're right. *But it becomes a rather tautological statement. |
therefore, once you commit that there can be change, you must accept
that there can be things that can change. If there can be things that
can change, but the system has no way to track identity across
changes, then the system is broken.
Quote:
|
For things that can change to be identical, the loci in time or space during which they exist or existed must coincide, and all changes in appearance that has been sustained by any at any time must have been sustained by all at that same time. *There can be no discernible difference between them at any time during their lifetimes. I feel you keep confusing things and statements about things. |
Quote:
|
A relational database records statements, usually about things. |
assigned something--some...thing--under an interpretation.
Quote:
|
-- Reinier |
#13
| |||||
| |||||
|
|
I am afraid I do not see the difference. I wrote that reply to ask you what the difference is. |
criticism is always welcome. Your comments, however, are more often
intended as degrading than constructive.
So I will reply on a lighter note.
Quote:
|
Second, it seems to me that your use of identifiers can be eliminated systematically and *losslessly*, i.e. without diminishing expressive power. *I'm curious whether this is the case. |
Example1. You need to make a db for a car dealer , who sells new
Honda Civic vehicles. All of these vehicles are identical. If you
eliminate the identifier, that is the VIN (vehicle identification
number) it will be impossible to make a db (in the RM) for this
dealer.
Example2. If you eliminate identifiers from any database, I suggest
the following algorithm:
Case1 - The key has less than 100 attributes.
If the server looks like Christmas tree, that is, if all
the lamps are blinking,
feel free to continue to do what you have been doing.
Case2 - The key has between 100 and 200 attributes.
If you notice smoke caming from server, consider calling
fire department.
Case3 - The key has more than 200 attributes.
Now buy a candle, because the fuse on the nearest power
plant will blow.
Note that there have been cases in this user group with relations that
have 200+ attributes.
In real life 99 procent of business applications use identifiers. This
happenms because IT professionals are ahead of IT theoretics. In my
paper, you can find a theoretical background for identifiers.
Quote:
|
You represent these 'meta-data' in a separate table and then link them to the 'concept' using an identifier attribute. |
please do not use them as mine, and then criticize me about them.
I do not think that time or people are meta-data. It would do our
discussion good if you gave us a drfinition for meta-data and data.
Otherwise, I will not take you seriously. ( You can find a definition
of data in my paper).
Quote:
|
I don't understand. *Suppose your database has an entity Person. I am a person. *What is my identifier? |
your identification and representation.
Quote:
|
Reinier |
#14
| ||||||
| ||||||
|
|
On 20 srp, 23:09, rp (AT) raampje (DOT) (none) (Reinier Post) wrote: I am afraid I do not see the difference. I wrote that reply to ask you what the difference is. Given that I have spent much time working on this paper, constructive criticism is always welcome. Your comments, however, are more often intended as degrading than constructive. |
relationship between your proposed model and the relational model.
It seems to me that it can be formalized in such a way that
your model can be thought of as a way of using existing DBMSes,
which would save a lot of work trying to understand or implement it.
Quote:
|
So I will reply on a lighter note. Second, it seems to me that your use of identifiers can be eliminated systematically and *losslessly*, i.e. without diminishing expressive power. *I'm curious whether this is the case. No this is not the case, I will give you two examples: Example1. You need to make a db for a car dealer , who sells new Honda Civic vehicles. All of these vehicles are identical. If you eliminate the identifier, that is the VIN (vehicle identification number) it will be impossible to make a db (in the RM) for this dealer. |
It is not something that requires you to go beyond the relational model.
[...]
Quote:
|
In real life 99 procent of business applications use identifiers. This happenms because IT professionals are ahead of IT theoretics. In my paper, you can find a theoretical background for identifiers. |
that identifiers are implementation details, funbdamentally wrong. etc.
I am not even one of those people. I'm just asking how you prove them wrong.
Quote:
|
You represent these 'meta-data' in a separate table and then link them to the 'concept' using an identifier attribute. Since I have not mentioned the terms meta-data or table in my paper, please do not use them as mine, and then criticize me about them. I do not think that time or people are meta-data. It would do our discussion good if you gave us a drfinition for meta-data and data. |
or retracted by a database change. Metadata: facts regarding such assertions
or retractions, e.g. when they are made, or by whom. The distinction is
necessarily interpretation dependent.
Quote:
|
Otherwise, I will not take you seriously. ( You can find a definition of data in my paper). |
Quote:
|
I don't understand. *Suppose your database has an entity Person. I am a person. *What is my identifier? You can try to follow your ideas and use your attributes strictly for your identification and representation. |
--
Reinier
#15
| |||||
| |||||
|
|
On Jul 27, 7:46*pm, rp (AT) raampje (DOT) (none) (Reinier Post) wrote: I wrote: [...] if you have no way to track identity across changes in real life, adding it as a modeling feature (either with explicit identities or by distinguishing between updates and deletes+inserts, as Brian proposes) isn't going to help a bit. Brian replies: [...] * Either there can be change, which implies that there can be things that can change, or there cannot be change, which means that there cannot be deletes or inserts, let alone updates. No, Brian. *These deletes, inserts, and updates are about statements of fact about the world, which can change to reflect new or changed observations, even when we haven't identified any objects that these statements are about to the extent you appear to deem necessary. *I can observe Mary's goldfish and its medical condition, and truthfully record that in my database, two days in a row, *regardless of* whether I can tell whether we're dealing with the same goldfish in both cases. Either there can be change, which implies that there can be things that can change, [...] |
I am not Parmenides. (Although Parmenides would doubt that.)
Quote:
|
Your argument neither affirms or denies my statement. What it does is illustrate the fact that descriptions, such as "Mary's goldfish," can refer to different things at different times in the same way that a particular instance of a composite key can refer to different things at different times--something I've been arguing for a long time. |
different things? My claim is that what it means you have a
*different* model in your mind in which the same thing is
represented only once. In other words, you are essentially
comparing two different models of the same reality, rather than
comparing a model to 'reality itself', something other than a model.
This is the mistake pretty much everybody I've ever been pointed to
regarding the issue of identity appears to make: they don't realize
it's essentially a language feature, and relative to the conceptual model
you're using, even if that model is incomplete or informal.
Quote:
|
Because interpreters are temporal beings, interpretations must occur / at a time/, which requires that there must be different interpretations at different times. Even under the possible worlds paradigm, only one possible world at a time can be the actual world. |
[...]
Quote:
|
[...] If there can be things that can change, but the system has no way to track identity across changes, then the system is broken. |
And we need to settle for compromises if we are ever going to settle
on a finite conceptual model of anything.
[...]
Quote:
|
Statements are /always/ about things. Each term in a statement is assigned something--some...thing--under an interpretation. |
make it true. If you stick with it, and fail to recognizing that identity
is a feature in a modeling language used to describe the world, rather
than the world itself, how are you ever going to get rid of those famous
'logical paradoxes' related to identity?
--
Reinier
|
« Previous Thread
|
Next Thread »
| Thread Tools | |
| Display Modes | |
Linear Mode
| |
Powered by vBulletin Version 3.5.3
Copyright ©2000 - 2012, Jelsoft Enterprises Ltd.
Copyright ©2000 - 2012, Jelsoft Enterprises Ltd.