Re: Orthogonal Was: Relational vs network vs hierarchic databases

"Laconic2" <laconic2_at_comcast.net> wrote in message news:FtednSAgVtvOuwjcRVn-jg_at_comcast.com...
>
> "Dan" <guntermann_at_verizon.com> wrote in message
> news:8x6ld.14$m36.6_at_trnddc02...
>>
>> "Dan" <guntermann_at_verizon.com> wrote in message
>> news:7a6ld.253$h15.215_at_trnddc07...
>> > >
>> >>>What does orthogonal mean? Does it mean that the intersecting vectors
> of
>> >>>relational and performance never ever really intersect? No.
>> >>
>
> I had thought that two vectors are orthogonal if and only if their dot
> product is zero(vector).

I don't think we are that far from each other in terms of definition really. By the above, you are just saying that two vectors are perpendicular, aren't you?

>
> This sounds like a somewhat more restrictive definition than one the you
> cited in your long article.

Well, probably not. When discussing the ISA instruction set and memory hierarchy, my point was really that at such a low level of implementation (or modeling), pure orthoganality is really much tougher and it might be useful to think in terms of degrees of orthoganality. But it was also to make the point that though independent to the greatest degree possible, there is still a sliver of relatedness and dependence and possibly either global or embedded knowledge on one side or the other. This is in contrast to "total independence".

In layered architectures, this point of convergence naturally falls on the interface.

>
> Is the example I cited above wrong? Am I misunderstanding your definition
> (in mathematics)?

I don't think so. But having reflected on it a lot after encountering such a grand misuse and overuse of the term, I've come to my own conclusions that the mathematical definition and the computer science definition are not necessarily different nor incompatible.

In computer science, my introduction to the term
> "orthogonal" comes from the description of several feature of Pascal:
> thus
> records and arrays are both compound types, but they are orthogonal
> meaning
> their utility has minimal overlap.

Yes. This sounds reasonable to me. The point being that if they didn't overlap in some way, they would be called "mutually exclusive" or "unrelated" or "totally independent".
>
> I really am asking for education, and not trying to nit-pick with you.
>
I understand entirely, but I'm fine with nit-picking as well. It seems like you need for little education from me, anyway.

Try this link and scroll down to the "Derived Meanings" section: http://en.wikipedia.org/wiki/Orthogonal. It seems incomplete, but what is written there doesn't contradict any of my own thoughts.

HTH,

• Dan