This article seems interesting:

http://www.readwriteweb.com/archives...ase_doomed.php

Is ingres doing anything to challenge cloud computing?

Chenzhen

zhenchen17 wrote:
[Standing by for the "Wrath of Roy"]

***
Mark R. Winston
***

"Mark R. Winston" <winstonmr (AT) datavailable (DOT) com> wrote

Roy has been assimulated. Resistance is futile.

zhenchen17 wrote:

Maybe it seems interesting, but it is ill-informed, so it actually
fails to make a number of very good points that it could have made.

It claims to talk about relational databases but then makes its points
with reference to SQL *products*, which are not at all the same thing.

Relational databases are not doomed and never will be, for the simple
reason that relational theory is just first-order predicate logic and
set theory applied to data management. Predicate logic has been around
since the ancient Greeks, so unless someone introduces Logic
2.0 soon, that is pretty secure. Set Theory has been around since the
late nineteenth century and is also pretty secure. There is not,
never has been, and is not likely ever to be a simpler, more regular,
more orthogonal way to describe data manipulation than relational
algebra/calculus. Anything that has ever claimed to be "better than
relational" has invariably turned out to be more complex and more
awkward, and always will.

So relational databases are not doomed.

What might be doomed is current SQL implementations. In particular the
implementation of a table as a "heap of records" in a file (or table
space). There is no reason for implementing tables like that, and very
good reasons not to. It was done that way because it was easy in the
late 1970s and we've been stuck there ever since. In fact there is no
good reason not to store data in any configuration (or indeed in
multiple configurations) that benefit performance, and that includes
instantiations of objects (in the object-oriented sense of the term).
As long at the DBMS knows how to present a tabular interface to the
underlying tangle so that you can use relational operators on it, that
is all that matters.

BTW, cloud computing may be what grabs the attention of the author of
the above-referenced paper, but what is really going to overthrow the
current table-as-a-heap-of-records mentality is massive solid state
memory. People have been predicting the imminent demise of the disk
drive for as long as I've been in this business (over 30 years), but it
looks like it is finally going to happen.

There are a number of initiatives that I am aware of. All I can suggest
is that you set up some Google Alerts for Ingres related stories.

--
Roy

UK Ingres User Association Conference 2009 will be on Tuesday June 9, 2009
Go to http://www.iua.org.uk/join to get on the mailing list.

On Web wrote:


Not so fast there bub! :-) I posted my "wrath" a minute ago.

--
Roy

UK Ingres User Association Conference 2009 will be on Tuesday June 9, 2009
Go to http://www.iua.org.uk/join to get on the mailing list.

On Feb 24, 1:26*pm, zhenchen17 <zhenche... (AT) gmail (DOT) com> wrote:
Sounds like a great business to be in if you're charging to host these
databases based on disk space... Lots and lots of yummy duplicated
data = megabucks... It also seems like a lot of hard work if a single
value needs amending that would normally only exist in one lookup
table in a well-designed relational database.

On Feb 24, 2009, at 8:26 AM, zhenchen17 wrote:

But alas, it turns out not to be.

My interest started to wane as soon as the author confused "database"
with
"database management system" as early as the seventh word. It failed
entirely with the second sentence, which is a foolish leap of
illogic that dwarfs anything Doctor McCoy ever did. ("Recently,
many red birds have been sighted, both on the ground and in the
air. One key message this sends is that if you want to fly and
arrive on time, you must take a plane painted red." Gah.)

I did make it into the second page before I decided to quit wasting
my time.

It's clear to me that the question is not "Is the relational
database [sic] doomed?" The real question that the author was
trying to answer was "My mortgage payment is due, how fast can
I sell a few words?"

Actually I think he missed the boat. Hand-wringing over the
putative demise of the relational database is passe. If he had
really wanted to attract some attention, he should have written
a piece on "Is XML Dead?" (Heck, why not go for broke and write
"Is the XML Database Dead?" -- it's not really any sillier.)
He might have had to change a word or two here and there, but
he could have substituted "XML" for "relational" and ended up
with a much trendier article.

Roy's point about SQL and relational not being equivalent is
well taken. Unfortunately, writing the article from that
perspective would have taken more depth of understanding than
I think either the author or most of the commenters have.

(I was especially amused by the first commenter, who was foolish
enough to propose hierarchical models replacing the relational
model! As if we haven't been there and done that.)

Cloud computing is one of those wonderful Humpty-Dumpty
phrases that means whatever the speaker wants it to mean.
So, I don't know. If we are fundamentally talking about
scalability, then yes, I do think that there is some work
going on in various back rooms, but I'm not sure exactly
what will come out.

Karl

At least he said basically "ignore me, I just typed some BS"...

my favorite was:
#10 "Almost all open source RDMS have a RAID volume as a central
spinning storage, maybe with seperate log and transaction volumes"

Well, I will admit that RAID is *usually* somewhere in the mix, but I
can't say as I've ever seen a DBMS, open source or otherwise, that
demanded that it run on RAID volumes...

Just as well he didn't actually say that RDBMS demanded RAID volumes.

Gosh the ingres group is insular.

On Feb 24, 2009, at 3:01 PM, On Web wrote:

We may be, but it certainly doesn't follow from this exchange as I've
read it so far. There are a few people who work with other DBMS's,
who understand things like the difference between "database" and
"DBMS", or what the relational model really says. :-)

I admit that the former is becoming largely a matter of precision
vs (sloppy) jargon, but the latter is kinda important.

Karl