"Marshall Spight" <mspight@dnai.com> wrote in message
news:5VBnd.361746$wV.199752@attbi_s54...
> "Laconic2" <laconic2@comcast.net> wrote in message
> news:QaWdnbbDXp-LjAzcRVn-3g@comcast.com... 

This is OT.  Sorry...
[vbcol=seagreen]
> I was a fairly staunch OO guy once upon a time, and I was rather
> convinced that relational databases were boring and stupid.

Out of curiosity, what was the impetus for such a broadening leap?  It seems
to me that individuals such as you are a rare breed indeed.  Of course
psychology, creative writing, and computer science are pretty broad and
independent strokes themselves.

> I'm not sure if any solid performance argument is necessary to
> motivate a verbal dismissal of relational.

Forget of course that that OO, particularly Java, is not the screaming
performance devil itself.  With the same logic in isolation, we could
advocate dumping OO and going back to assembly language (or even C).  So
perhaps OO proponents should know better...

> Marshall
>

- Dan

[ Post a follow-up to this message ]

"Dan" <guntermann@verizon.com> wrote in message
news:HHCnd.1160$0k1.504@trnddc08...
>
> "Marshall Spight" <mspight@dnai.com> wrote in message
> news:5VBnd.361746$wV.199752@attbi_s54... 
>
> This is OT.  Sorry...

And this is OT off of your OT.

I'd say I've had one broadening leap about every ten years in this
profession.

They are generally "behind the times".  There's a reason for that.  To me,
going from programming to database was a broadening leap.  The chioce of
relational (later SQL) databases over CODASYL databases was one I made for
very simlpeminded reasons:  it's easier to get something done.

[ Post a follow-up to this message ]

"Dan" <guntermann@verizon.com> wrote in message news:HHCnd.1160$0k1.504@trnddc08...
>
> "Marshall Spight" <mspight@dnai.com> wrote in message
> news:5VBnd.361746$wV.199752@attbi_s54...
> 
>
> Out of curiosity, what was the impetus for such a broadening leap?  It see
ms
> to me that individuals such as you are a rare breed indeed.  Of course
> psychology, creative writing, and computer science are pretty broad and
> independent strokes themselves.

Well, gosh, what a flattering question.

I had been working on middleware for a few years. I'd gotten a
job (via my OO skills) programming ODBC, OCI, etc. drivers.
I didn't really know the API very well or what it was used for,
but you can network-enable anything without actually knowing
much about it.

I recall a conversation with a PM (no less) in which I had proposed
some dumb idea of how to model something with tables. The PM
said, you'd more typically do some things like this. And he showed
me a many-to-many table, and how you could use it and query it,
etc. And suddenly these seemingly-simple relational operators
were doing something quite clever, and it was actually doing
something that was problematic in OO.

So after having mocked SQL for some years, I made it a point to
learn more about it. I had heard this term "third normal form" and
started doing web searches for it, and went and bought some books
by Date.


> Forget of course that that OO, particularly Java, is not the screaming
> performance devil itself.  With the same logic in isolation, we could
> advocate dumping OO and going back to assembly language (or even C).  So
> perhaps OO proponents should know better...

I dunno. I just don't care that much about performance any more.
I mean, I care if something is O(n^2) vs. O(n log n) but I don't
so much care about language speed. Java is fast enough.

Actually, it's funny what perspective brings. I can recall a whole
host of arguments (many performance related) that the C world
tried to use to shoot down C++, and then, ten years later, many
of those same style of arguments being fired at Java, *by the C++
people.*


Marshall

[ Post a follow-up to this message ]

"Marshall Spight" <mspight@dnai.com> wrote in message
news:XkUnd.58923$V41.58239@attbi_s52...

> I recall a conversation with a PM (no less) in which I had proposed
> some dumb idea of how to model something with tables. The PM
> said, you'd more typically do some things like this. And he showed
> me a many-to-many table, and how you could use it and query it,
> etc. And suddenly these seemingly-simple relational operators
> were doing something quite clever, and it was actually doing
> something that was problematic in OO.

I think that this reveals something.  It's the join operator that makes the
system of relations so powerful.
The restrict (select) operator is a pretty natural revision of "record
selection criteria",  and the project is not easily appreciated at first.
But it's the operators that make relations useful as a type.

[ Post a follow-up to this message ]

"Marshall Spight" <mspight@dnai.com> wrote in message
news:XkUnd.58923$V41.58239@attbi_s52...
> "Dan" <guntermann@verizon.com> wrote in message
news:HHCnd.1160$0k1.504@trnddc08... 
seems[vbcol=seagreen] 
>
> Well, gosh, what a flattering question.
>
> I had been working on middleware for a few years. I'd gotten a
> job (via my OO skills) programming ODBC, OCI, etc. drivers.
> I didn't really know the API very well or what it was used for,
> but you can network-enable anything without actually knowing
> much about it.

I think it a little interesting that I went the other direction -- from
RDBMS knowledge to OO and Java (a little late on the scene, having learned
Java in the last three years).  They say that "converts" are the biggest
evangelists.

> I recall a conversation with a PM (no less) in which I had proposed
> some dumb idea of how to model something with tables. The PM
> said, you'd more typically do some things like this. And he showed
> me a many-to-many table, and how you could use it and query it,
> etc. And suddenly these seemingly-simple relational operators
> were doing something quite clever, and it was actually doing
> something that was problematic in OO.
>
> So after having mocked SQL for some years, I made it a point to
> learn more about it. I had heard this term "third normal form" and
> started doing web searches for it, and went and bought some books
> by Date.
>
> 
>
> I dunno. I just don't care that much about performance any more.
> I mean, I care if something is O(n^2) vs. O(n log n) but I don't
> so much care about language speed. Java is fast enough.

I absolutely agree.

> Actually, it's funny what perspective brings. I can recall a whole
> host of arguments (many performance related) that the C world
> tried to use to shoot down C++, and then, ten years later, many
> of those same style of arguments being fired at Java, *by the C++
> people.*

It is called "change" -- we heard the same things from the assemby language
programmers who were forced to code in COBOL (and tended to write the worst
COBOL code, grabbing memory locations 'cause they could, for example).
Cheers!  --dawn

>
> Marshall

[ Post a follow-up to this message ]

"Dan" <guntermann@verizon.com> wrote in message
news:HHCnd.1160$0k1.504@trnddc08...
>
> "Marshall Spight" <mspight@dnai.com> wrote in message
> news:5VBnd.361746$wV.199752@attbi_s54... 
>
> This is OT.  Sorry...

And this is OT off of your OT.

I'd say I've had one broadening leap about every ten years in this
profession.

They are generally "behind the times".  There's a reason for that.  To me,
going from programming to database was a broadening leap.  The chioce of
relational (later SQL) databases over CODASYL databases was one I made for
very simlpeminded reasons:  it's easier to get something done.

[ Post a follow-up to this message ]

"Dan" <guntermann@verizon.com> wrote in message news:HHCnd.1160$0k1.504@trnddc08...
>
> "Marshall Spight" <mspight@dnai.com> wrote in message
> news:5VBnd.361746$wV.199752@attbi_s54...
> 
>
> Out of curiosity, what was the impetus for such a broadening leap?  It see
ms
> to me that individuals such as you are a rare breed indeed.  Of course
> psychology, creative writing, and computer science are pretty broad and
> independent strokes themselves.

Well, gosh, what a flattering question.

I had been working on middleware for a few years. I'd gotten a
job (via my OO skills) programming ODBC, OCI, etc. drivers.
I didn't really know the API very well or what it was used for,
but you can network-enable anything without actually knowing
much about it.

I recall a conversation with a PM (no less) in which I had proposed
some dumb idea of how to model something with tables. The PM
said, you'd more typically do some things like this. And he showed
me a many-to-many table, and how you could use it and query it,
etc. And suddenly these seemingly-simple relational operators
were doing something quite clever, and it was actually doing
something that was problematic in OO.

So after having mocked SQL for some years, I made it a point to
learn more about it. I had heard this term "third normal form" and
started doing web searches for it, and went and bought some books
by Date.


> Forget of course that that OO, particularly Java, is not the screaming
> performance devil itself.  With the same logic in isolation, we could
> advocate dumping OO and going back to assembly language (or even C).  So
> perhaps OO proponents should know better...

I dunno. I just don't care that much about performance any more.
I mean, I care if something is O(n^2) vs. O(n log n) but I don't
so much care about language speed. Java is fast enough.

Actually, it's funny what perspective brings. I can recall a whole
host of arguments (many performance related) that the C world
tried to use to shoot down C++, and then, ten years later, many
of those same style of arguments being fired at Java, *by the C++
people.*


Marshall

[ Post a follow-up to this message ]

"Marshall Spight" <mspight@dnai.com> wrote in message
news:XkUnd.58923$V41.58239@attbi_s52...

> I recall a conversation with a PM (no less) in which I had proposed
> some dumb idea of how to model something with tables. The PM
> said, you'd more typically do some things like this. And he showed
> me a many-to-many table, and how you could use it and query it,
> etc. And suddenly these seemingly-simple relational operators
> were doing something quite clever, and it was actually doing
> something that was problematic in OO.

I think that this reveals something.  It's the join operator that makes the
system of relations so powerful.
The restrict (select) operator is a pretty natural revision of "record
selection criteria",  and the project is not easily appreciated at first.
But it's the operators that make relations useful as a type.

[ Post a follow-up to this message ]

"Marshall Spight" <mspight@dnai.com> wrote in message
news:XkUnd.58923$V41.58239@attbi_s52...
> "Dan" <guntermann@verizon.com> wrote in message
news:HHCnd.1160$0k1.504@trnddc08... 
seems[vbcol=seagreen] 
>
> Well, gosh, what a flattering question.
>
> I had been working on middleware for a few years. I'd gotten a
> job (via my OO skills) programming ODBC, OCI, etc. drivers.
> I didn't really know the API very well or what it was used for,
> but you can network-enable anything without actually knowing
> much about it.

I think it a little interesting that I went the other direction -- from
RDBMS knowledge to OO and Java (a little late on the scene, having learned
Java in the last three years).  They say that "converts" are the biggest
evangelists.

> I recall a conversation with a PM (no less) in which I had proposed
> some dumb idea of how to model something with tables. The PM
> said, you'd more typically do some things like this. And he showed
> me a many-to-many table, and how you could use it and query it,
> etc. And suddenly these seemingly-simple relational operators
> were doing something quite clever, and it was actually doing
> something that was problematic in OO.
>
> So after having mocked SQL for some years, I made it a point to
> learn more about it. I had heard this term "third normal form" and
> started doing web searches for it, and went and bought some books
> by Date.
>
> 
>
> I dunno. I just don't care that much about performance any more.
> I mean, I care if something is O(n^2) vs. O(n log n) but I don't
> so much care about language speed. Java is fast enough.

I absolutely agree.

> Actually, it's funny what perspective brings. I can recall a whole
> host of arguments (many performance related) that the C world
> tried to use to shoot down C++, and then, ten years later, many
> of those same style of arguments being fired at Java, *by the C++
> people.*

It is called "change" -- we heard the same things from the assemby language
programmers who were forced to code in COBOL (and tended to write the worst
COBOL code, grabbing memory locations 'cause they could, for example).
Cheers!  --dawn

>
> Marshall

[ Post a follow-up to this message ]

"Dan" <guntermann@verizon.com> wrote in message news:HHCnd.1160$0k1.504@trnddc08...
>
> "Marshall Spight" <mspight@dnai.com> wrote in message
> news:5VBnd.361746$wV.199752@attbi_s54...
> 
>
> Out of curiosity, what was the impetus for such a broadening leap?  It see
ms
> to me that individuals such as you are a rare breed indeed.  Of course
> psychology, creative writing, and computer science are pretty broad and
> independent strokes themselves.

Well, gosh, what a flattering question.

I had been working on middleware for a few years. I'd gotten a
job (via my OO skills) programming ODBC, OCI, etc. drivers.
I didn't really know the API very well or what it was used for,
but you can network-enable anything without actually knowing
much about it.

I recall a conversation with a PM (no less) in which I had proposed
some dumb idea of how to model something with tables. The PM
said, you'd more typically do some things like this. And he showed
me a many-to-many table, and how you could use it and query it,
etc. And suddenly these seemingly-simple relational operators
were doing something quite clever, and it was actually doing
something that was problematic in OO.

So after having mocked SQL for some years, I made it a point to
learn more about it. I had heard this term "third normal form" and
started doing web searches for it, and went and bought some books
by Date.


> Forget of course that that OO, particularly Java, is not the screaming
> performance devil itself.  With the same logic in isolation, we could
> advocate dumping OO and going back to assembly language (or even C).  So
> perhaps OO proponents should know better...

I dunno. I just don't care that much about performance any more.
I mean, I care if something is O(n^2) vs. O(n log n) but I don't
so much care about language speed. Java is fast enough.

Actually, it's funny what perspective brings. I can recall a whole
host of arguments (many performance related) that the C world
tried to use to shoot down C++, and then, ten years later, many
of those same style of arguments being fired at Java, *by the C++
people.*


Marshall

[ Post a follow-up to this message ]