What so special about PostgreSQL and other RDBMS?

[Sarah Tanembaum]

Beside its an opensource and supported by community, what's the fundamental
differences between PostgreSQL and those high-price commercial database (and
some are bloated such as Oracle) from software giant such as Microsoft SQL
Server, Oracle, and Sybase?

Is PostgreSQL reliable enough to be used for high-end commercial
application? Thanks

[Erland Sommarskog]

Dmytri Kleiner (qu***@syntac.n et) writes:

Good thing that you only mislead a few customers into overpaying for
crap. Your company is just a bankruptcy waiting for a competent
competitor to make it happen.
Our customers seems to be quite satisfied with our system.

And - in difference to you - they actually know the system in question,
so I think they are somewhat better apt to tell whether it is crap or
not.
What bunk, saying the competitor tried 'precicely what I teach' and
thus failed is an obvious attempt to fallaciously discredit my
argument with out actually addressing it. You're a ham fisted shill.

I think that I made it quite clear in my first post that your suggested
strategy indeed may be very valid sometimes. But what I've been pointing
out is that this far from always the case.

Doing the sort of abstraction you suggested is *very* expensive, and
for small companies like ours or our competitor, this a huge enterprise
to take on for systems with over 500 tables and over 3500 stored procedures.
(That data is for our system; Obviously I don't have the data for our
competitor's system, but I do know the business they were targeting.)

--
Erland Sommarskog, SQL Server MVP, so****@algonet. se

Books Online for SQL Server SP3 at
http://www.microsoft.com/sql/techinf...2000/books.asp

[Quirk]

Erland Sommarskog <so****@algonet .se> wrote in message news:<Xn******* *************** @127.0.0.1>...

Dmytri Kleiner (qu***@syntac.n et) writes:

Good thing that you only mislead a few customers into overpaying for
crap. Your company is just a bankruptcy waiting for a competent
competitor to make it happen. Our customers seems to be quite satisfied with our system.

And - in difference to you - they actually know the system in question,
so I think they are somewhat better apt to tell whether it is crap or
not.
The Application that you wrote I have no reason to doubt is of
sufficient quality to keep your customers satisfied.

Unfortunately you have created unneeded dependencies for them, the
worst of which is not MS SQL, since it is fairly easy to get at data
in MS SQL and archive it or export it in a usefull way, the worst is
that you have tied your customers to a terrible Operating System with
a terrible licence, even Oracle users are not so screwed since at the
very least they have a choice when it comes to OS.
What bunk, saying the competitor tried 'precicely what I teach' and
thus failed is an obvious attempt to fallaciously discredit my
argument with out actually addressing it. You're a ham fisted shill.

I think that I made it quite clear in my first post that your suggested
strategy indeed may be very valid sometimes. But what I've been pointing
out is that this far from always the case.

I would say you made it quite clear that your basic message was that
it would be folly to do what I was suggesting, and that was your whole
purpose in posting, as I said, I can tell this by the obvious
rhetorical devices you used, claiming this unnamed third party did
'precicely what I teach' a wretchedly unlikely and unqualified
generalisation, followed by saying that this, implying that this
_alone_, lead to the failure of their project. This is obvious FUD.
The lack of any other content, or even specifics in your post is the
final damning evidence. As I said, you are a shill, and a ham fisted
one at that.
Doing the sort of abstraction you suggested is *very* expensive,
It is not, as I've said, it can be as simple as writing a wrapper
function around your data access.

Not as expensive as having the system itself obsoleted by an obsoleted
dependency or the inabilty to get support for a dependency due to a
licencing dispute.

But for you, this is probably useless advice, since no doubt not only
have you chosen a SQL server with a bad licence, and an OS with a bad
licence, but no doubt you have also choses a development platform with
a bad licence, let me guess: Visual Basic?

As I said, enjoy your solvency while it lasts.
and
for small companies like ours or our competitor, this a huge enterprise
to take on for systems with over 500 tables and over 3500 stored procedures.
(That data is for our system; Obviously I don't have the data for our
competitor's system, but I do know the business they were targeting.)

All the more reason to protect your investment and that of your
customer by not getting trapped into becoming dependant on a third
party for the continued operation of their own system.

But it's pretty clear that encouraging such pitiable dependencies is
exactly what you are here to do.

[Steve]

Daniel Morgan <da******@x.was hington.edu> wrote in message news:<108369595 7.163784@yasure >...

Sarah Tanembaum wrote:

Beside its an opensource and supported by community, what's the fundamental
differences between PostgreSQL and those high-price commercial database (and
some are bloated such as Oracle) from software giant such as Microsoft SQL
Server, Oracle, and Sybase?

Is PostgreSQL reliable enough to be used for high-end commercial
application? Thanks

PostgreSQL is highly overrated and not suitable for any environment
where little things like crash recovery and security are a priority.

What database does Google use?

Steve

[Volker Hetzer]

"Quirk" <qu***@syntac.n et> schrieb im Newsbeitrag news:4e******** *************** **@posting.goog le.com...

"Volker Hetzer" <vo***********@ ieee.org> wrote in message news:<c7******* ***@nntp.fujits u-siemens.com>...

> > > That's not true.

> > Yes it is.

> What was the value of this reply?

What was the value of yours? Or this latest one?

A question is not an answer.

And what was your reply?

I asked first.

Yes, you have the right to be overcharged for work that may or may not
not suit your needs by only _one_ vendor, and no right to go elsewhere
when they fail, ignore you outright, stop supporting your application
or vanish from the face of the earth. Have you actually read your
contract or software licence? Of course. See the end of this posting.

It only protects the vendor, not you.

I've read the licence and done even more: I've used the software and tested the contract.

Realy, care to quote the part of the Contract that Gaurantees you any
rights?

http://oracle.com/support/index.html?policies.html
By "tested the contarct" what you mean is you agreed to pay them
completely on their terms and where satisified with the results they
chose to give you. So, in what way is it different from let's say, buying a cucumber?

Have you tested alternatives? The other example was buyig gcc support from cygnus.
One bug, never got resolved in one year, consequently
we cancelled support.

> > > The right to modify is a red herring.
> >
> > Not if your application and the permenancy of your data is important.

> You didn't read my posting, right?

You are one funny guy. Really. I'll bet you're the first guy in usenet
to ever ask this question rhetoricly.

Nice way of avoiding an answer.

Are so so stupid that you actually expect a serious answer that was
obviously a
hostile attempt to insult by way of a rhetorical question?

Ok, so for you explicitly: That was not a rhetorical question. Your response
indicated youy didn't read my posting, or at least not the relevant part, so I
wanted to check whether it was worth posting any more.

I don't *want* to create my own development
> team competing with the original one. I don't want to merge my change back
> into their code with every new release! I don't want to develop code and
> then have them decide whether they condescend to incorporate it or not! I
> want the authors of the software to do the coding based on what I'm willing
> to pay for!

You are dependent on their licence

I'm dependent on the author's licence regardless of which database I use.
Yes, which is why you should choose one that give you a perpetual
right to the source code, otherwise you are locked into a dependancy
that may prove fatal to your application.

I start to repeat myself here. The right to the source code does not mean
anything useful, see the part you quoted below.

It's just that some licences give me the illusion of being able to do
something while mainly giving me in reality the ability to shoot myself in
the foot or paying someone else to shoot me in the foot.
Unsubstantiated bunk, if you have the source code, it is not magic to
fix it, or extend it, just normal progamming.

Right. So, if I do CAD programming, why should I learn database programming
only to support a dead database? It's much easier to migrate to another one.

Besides, have you considered that quite a few open source projects get abandoned
because they have become unmaintainable? Anyone remembers hurd? Groff?
What was the last gmake improvement? And if the authors throw up their hands,
what can I do? Ask my boss to form a department for the beating of dead horses?
Simple calling something
an illusion does not explain why you condsider it impossible to
actually change a program. Perhaps you should consider a different
line of work. Oh, it's pretty easy to change a program. Working through millions
of lines of code and repairing it with less time or money than it would
cost to migrate to another database is the trick. Convincing the customer to
install *my* database version is another, particularly if three or four
developers do this.
because you built your own
application on top of a platform for which you have no source code,

Same question: Did you read what I wrote?
A better question: What kind of an idiot are you that, in the face of
good sense, the best you can do is attemp insulting, evasive
rehetoric?

It's not a better question. You keep bringing up that stupid
source code argument totally ignoring the fact that it simply doesn't
work, at least not for the money a normal support contract costs.
And if support doesn't work, I still won't support it on my own.

I don't care about the source code, I care about product and support
quality. And, since I am not the developer of the software, nor is anyone else,
apart from *the* developers,
As I said, my comments where ment *FOR DEVELOPERS* that is those who
are developing *NEW* appliciations, and my advice is simple enough,
despite your contortions: If your application is important to you, do
not engineer a dependency on code you do not have access to.'

Do you develop for platforms other than linux?

anyone else is going to make a worse job than
them. So, I get the best support when I'm paying them and no one else.
More unsubstantiated bunk, first of all, in many cases you can hire
the original developers,

Yeah, exactly. A man year here costs about USD200000,-. A support
contract with oracle costs me about a tenth of that.
And even if I buy some incident based support contract, there is still
no difference from an incident based support contract with oracle.
As long as that guy exists and I can sue him into doing his job I don't
need the source code (he needs) and otherwise I have no one to
replace him.
But thanks for acknowleding that reliable support costs money.
regardless of your right to the source code,
secondly, by hiring the "Copyright Holders" you *ARE NOT NECESSARLIY
HIRING THE DEVELEORS*, who may not even be with the company anymore,
in fact you are often hiring some peon who they scooped of the
consulting market 5 minutes before sending him to your office as an
certified solutions prodiver or whatever idiotic buzzword whey have
for their unskilled labour. Try it. Besides, remember, the company has an interest in providing
support because they live off it.

And finaly, it is a falalcy to say that someone will do a worse job
simply because they are not the original developer. So, if I pick some average application programmer off the street,
how long do you think it takes before he can start smoothing
out bugs in the postgres optimizer?
and no right to modify, you then also have no leverage with the vendor
of the orginal software.

You have no rights at all, wether or not you are willing to pay. Read oracles licence some time. There it says very clearly what
you get if you enter a support agreement.

But it stops short of guaranting that your apllication will actualy
work,

Of course they don't offer that. But they offer to put effort
in it. And they are dependent from me for my money.
or that your existing version of the software will be supported. They provide upgrades and desupport dates. Ok, they do
what I pay for.

In anycase, I am not arguing agianst using Oracle, as I said, if
Oracle suits your needs and you think it's worth the money, use it,
however, my advice is that if you do develop an application, write
your code in such a way that you do not depend on Oracle, but can
easily switch it over the the greatest extent possible. Why "the greatest extent"? That costs me more time and money
and customers that it's worth. Just look at informix to see how
it goes when a db disappears from the market:
They had a big market share, market share dwindled, they got weak
and sold themselves to ibm because that's better than going bancrupt.
Now IBM handles the migration to db2 and supports me as application
developer in porting my app to db2. This is much better than handing
me the source code and telling me that from now on I have to develop
all the new features and fix bugs on my own or simply buy a new db
and do the migration on my own.

I have no idea why you are insisting on jumping up and down like this
is crazy talk, the only plausible theory is that you get some kind of
thrill out of embarassing yourself. Where do I jump up and down?

I dunno, because you're culturaly issolated and have a poor
imagination?

No, it's because the phrase "elegant coding" is just as empty.
Or as the phrase "the one true god" uttered by people of
different religions.
This is just stupid, elegnt coding is hardly as unatainable an ideal
as you seem to be conviced, in fact in this specific case it's a
simply matter of using a standard wrapper function throughtout your
aplication to access your data rather than using proprietary bindings
throughout your application, if your application is sufficently
complicated, perhaps a data abstaction object might be usefull for
this function, perhaps not, if you use any non standard features of
your database server, then write some additional functions as wrappers
for these. It is anything but rocket science.

So you have defined "elegant" as "abstractio n" and expect the rest
of the programmers to agree that that's it?
Thanks for solving that problem for the rest of the world.
For db computing, reducing server load is the important thing.

No, it is not, in most cases CPU is not the most limited resource.

> Interoperabilit y
> typically means primitive, network/db intensive sql. Yup. Which, in a well configured db is CPU load because
caching, indexing and db specific sql takes care of the i/o load.

What about the human and financial load? As in the load on the DBA,
inhouse developers, consulting budgets and application support staff?

The load on the DBA depends on the problems the application makes.
That typical increases if the application ignores load reducing features for the
sake of being generic. This creates an excessive amoung of simple
queries and lots of network traffic. Right now we have huge problems
getting an application to work properly that claims to support mysql and
oracle. They could have done half the app in PL/SQL and saved 90%
of the network and client load.
Also, if the database is not the standard one (because you have
fixed/improved it) I have, at the worst, maintain two independent installations,
at best, two independent update cycles.
Developers are constrained by (among other things) the load they are allowed
to put on the db. That's a business decision.
As for consulting, we pay a flatrate for db support, so we unload as much
of our problems on the oracle people. Works fine.
Ditto for support staff. Our users have oracle, so the more we make the db do
the less problems we have in our own code.

Nevertheless, I concede, it *is* possible to have such a
horribly configured system that i/o load becomes an issue. It's also
possible to have a database that permits so few actions
that the dba can't do anything about a badly written app.
fortunately, oracle is different.

No, interoperabilit y means abilty to integrate applications in a
heterogeneus environment. It means standards and flexibilty.

So? What's more "standardis ed" about mysql's socket interface than
about oracles OCI or ESQL?

Are you having a nightmare in which we are dicussing the various
merits of MySQL versus Oracle? Please follow your own advice and read
this thread again so that you might figure out what is it we are
actually taking about.

We are talking about open source versus commercial databases. I picked
those two as examples because I have worked with both of them.

> > > If it's important it must not matter whether one tries to
> > > access the data from a local or remote machine.

> > Interesting that you believe that this can not be accomblished with
> > network security.

> Yes. Now you figure out why.

Because you don't know what you are doing maybe?

Wrong. Try again.

The more you talk, the clearer it is how right I was.
Oh wait, you don't
need to, after all, you have decided to pay a vendor to know for you,
I remember now.

Right. The alternative is not paying anyone and trying to figuring out the
source code on my own, right? Or paying someone else who starts
from scratch too?

More straw men and red herrings. If you are a Developer, which is who
my comments are addressed to, it is your responsiblilty to your users
and clients to know how your application works and to be able to
support it without allowing some third party to hold them hostage.

No one holds anyone hostage. I let people do what they are good at.
I'm ok with application programming in the CAD world. Oracle (or
IBM, or microsoft) are good at programming databases. So, I
profit from their expertise by being able to provide a better application
than if I had to do db development (or fixing) as well.
So far no one has complained.

Care to elaborate? An insecure network does not mean that someone can
log on to the database server from anywhere but the console screwed onto
it. And securing the listener (in case of oracle) is part of the database
configuration.

If the above is true, that someone can only access any of the devices
on your Database server via the local console, then your network *IS*
secure

one can of course log on to the database. Via the listener and all the stuff.
In theory from an unsecure network. So, db security is not network security,
because all the stuff of protecting data from different users needs the
cooperation of the database.
This is what I'm trying to say, that network security comes first, You can have a secure database within an insecure network.
because Database security can only depend on it, not being able to
actualy protect devices, which is the burden on the OS and networking
environment. The os protects devices, not the network. Or, daring to think the unthinkable,
do you mean that you consider it ok to have database data on nfs mounts?

Once again, It must be assumed that your consternations to contend
this point are some weird form of self-flagilation.

> Right. You show me how do convert VENUS chip designs into Synopsys
> without going into a museom for the original hardware and getting all
> the versions in between.

What does this have to do with "Self Contained, Self Describing, Human
Readable" files that can be read on any system past or present?

It has to do with permanency. Try to read what you quote.
What does reading text files have to do with Chip design?

Because some tool will have to parse the text and create the chip out of it.
This tool typically costs in the range of USD100000-200000 for a synopsys
ASIC compiler. You need the same tool because any other tool creates
totally different designs, ignores the original constraints and rules and
uses a different library which may even force a complete redisign.
Compared to that, a database migration is truly a breeze.
I can read
text files I created on my Apple ][, and no, I do not have the orginal
hardware (well maybe my mom does somewhere in her basement). Not all textfiles are notices for you to read.

Try to avoid making an ass of yourself with further pretentions.
tell me, how much do you know about my
experience,

What your arguments tell me.
Which ones? That abstracting access to suspect dependencies is a good
idea?

That elegance is abstraction. That database security is secondary to network security? Yes. That
one should keep archives in a format that is likely to be readable
forever? Yes.
Those are the tree main reasons. The fourth one is your persistent
belief that the right to the source code is of value.

All these things come from experience, So, what migrations have you done so far?
Right now I'm in the process of doing two, one boing our board design
toolchain, with plenty of data translation and the other a business flow app.
So far we've spent at least four man years on the CAD flow and it's far
from over. As for the other, try to imagine having a small busines flow
tool and then introducing SAP companywide.
And we get migration support from the new vendor.
Believe me, a database migration is *EASY* compared to that.
Even if I hardwire OCI calls into my c-code and then switch to
ODBC or something.
your attempt to question my
experience, only show that you are unable to formalute an actual
argument, so you try and discredit the arguer instead of the argument. I did. You just didn't understand it.
and why do you feel that talking about _me_ is a response
to my argument?

Because your argument isn't backed by anything. Give me some
substance and we can talk about it.
Oh please, my argument has been presented well enough, attacking me
just shows you can not defend your own, that is if you actually had
one.

You might have noticed that you got responses from different people
whereas you are the only one who thinks my arguments are rubbish.
Now, statistics is not fact, but it's evidence and should get you thinking.

If my argument was not backed up by anything it would easy enough to
refute it without attempting to insult me, I don't insult you I'm trying to get through to you.
Reasonable arguments didn't work.

All I've hear so far is the
usual open source rethoric about me or someone else being able
to magically support a product in a few days or weeks after the
original developers have abandoned it, or me.

These must be voices in your head that you are hearing. Since my
argument have been quite clear and even sumerized several times.

Yes. The right to source code balances nonexisting support and
buying support for a open source software (instead of trying to
fix things oneself) is somehow better than doing the same with
commercial software. Did I leave out anything important?

Your arguments amount to the metaphysical belief that only the
copyright holders of your favourite proporiety software know how to
program, No, that they are the only ones that should be allowed because they
are the only ones that can take responsibility.
that the very concept of good programming is an illussion, No, it's just that so far no one has found out what it is, because
despite all the attempts software still is not substantially more stable
than software written 30 years ago.
and therfore the only way forward is to make yourself both tehnicaly
and legaly dependent on them as much as possible. You forget that they depend on me. Namely, on my money.
> If you have the source code, you are the developer,

> Wrong. I am the user, t.

Oh, well then I guess we have nothing further to discuss, my comments
here where meant for actual developers.

So, oracle people should further develop oracle and mysql people
mysql. Did I get this right?

No, that's not right, that's not even wrong.

So, what is it?

(with applogies to Wolfgang Pauli)

Application developers should avoid locking themselves in to external
dependencies, either by not using products to which they have no right
to the source code, or abstracting access when they do use such
products. Simple. There it is again, this source code right thingie. And you complain about me
getting rude.
Again: The source code is no guarantee of fixed bugs, much less improvements.
It's not even what I want. I also can go and tinker with the airbag of my car
if I think it's broken, I don't do that either but go to a repair shop.
And if you are worrying about expiring licences, for many products
(purify and our oracle installation spring to mind) you get permanent licences and pay yearly
for support, so I can still use the app when the vendor goes bust.
And before you come again about the source code I can fix and improve,
or pay someone to do it, I won't because that would be wasting company
money and that would be because a migration is cheaper than tinkering with
the old software and it wouldn't lose us customers either because customers
don't like dead software.
When we figured out that our new CAD tool doesn't support oracle 9.2
we gave them a ding behind the ear and, see, the next release, out
in two months supports it and til then we got a workaround.

And having right to the source code does not mean that the program is
'open source,' as you can purchace such a right for propretary code,
as is common for libraries. And still, if something goes wrong, I file a service request.
And if the company does ceases to offer the product I change company.

Of course, when the program _is_ open source, you are guaranteed that
right.
You have no such right, ever, the only right you _can_ have is the
right to modify it yourself or contract someone to do it. Please read
your licence.

"Assistance with my SRs 24 hours per day, 7days a week". Practically I
usually get two or three guys working on a typical SR of mine, depending on
how log it takes. Without a contract I'd get a 'buzz off, I'm doing my exams > this month'.

"Assitance" only means that they will provide someone whose time they
can bill you for,

As I said, we pay a flatrate.
not that anything will be accomplished. Then they lose money if they don't accomplish anything.
Many large companies, and profesional develpoers provide source
licences and/or support open source products, including the largest
computer company in the world, IBM.

Yep, so I can buy support, mess up the code I've access to and let
IBM sort it out, is this what I get by using a IBM supported mysql?
If not, what's the difference to buying db2 support?
(One thing more: No, if IBM abandons mysql I'm still not taking
on the support task, ok?)

Greetings!
Volker

[Geoff Berrow]

I noticed that Message-ID:
<6f************ **************@ posting.google. com> from Steve contained
the following:

PostgreSQL is highly overrated and not suitable for any environment
where little things like crash recovery and security are a priority.

What database does Google use?

Google's data is stored in data coops.
http://www.google.com/technology/pigeonrank.html

--
Geoff Berrow (put thecat out to email)
It's only Usenet, no one dies.
My opinions, not the committee's, mine.
Simple RFDs http://www.ckdog.co.uk/rfdmaker/

[Volker Hetzer]

"Quirk" <qu***@syntac.n et> schrieb im Newsbeitrag news:4e******** *************** **@posting.goog le.com...

"Volker Hetzer" <vo***********@ ieee.org> wrote in message news:<c7******* ***@nntp.fujits u-siemens.com>...

"Quirk" <qu***@syntac.n et> schrieb im Newsbeitrag news:4e******** *************** *@posting.googl e.com...

ed*****@highstr eam.net (Edward Lloyd Hillman) wrote in message news:<10******* ******@news.sup ernews.com>...

> > You have no such right, ever, the only right you _can_ have is the
> > right to modify it yourself or contract someone to do it. Please read
> > your licence.
>
>
> Got a news flash for ya...

Oh boy, it's Seseme Street News, OK Kermit, keep talking.

> If you have a maintenance contract with a vendor and something of
> theirs' is broken, they must fix it if you need it.

Perhaps, but when the product in question is proprietary you have no
recourse when they fail, because no one else has any right to modify
the source code.

I have, at most, the right to sue them,
What cold comfort that is. I would prefer the right to make my
aplication work without their good graces.

Before you consider suing them I suggest you reiview your contact with
an actual lawyer. So you can understand exactly how painted into a
corner you really are.

Look, we've got about 50 people here dealing with
exactly those questions, telling us what contracts to enter and what not.
When we buy support, we *know* what we are in for and when and
what to sue them for and how to deal with them before we sue them.

at least, the right to cancel the
contract which hurts them way more

How can you cancel the contract when your entire application is
dependanton there product? Can you afford to throw away your
application too?

See my other posting. Compared to changing the application (replacing
it with another), changing the underlying database is easy.

than if I go to a postgres developer
and tell him I'm not interested any more. So, unlike open source developers,
they actually have an interest in doing something.
What on earth are you trying to say here? Why is a postgresql
developer any more or less interested in your contarct than one who
pedals proprietary software?

A developer who does not earn mony by it is less interested in providing
work than one who does. Therefore, support contracts make sense.
I was talking about the case where I go to the developer and ask him
to do something for free.
When you have a right to the source code you can sign such a contarct
with any firm you like, and fire their ass and hire another when they
fail. But it doesn't make sense to use any other firm than the guys who wrote it.

Why? What magic powers are possesed by the firm that holds the
copyright? Expcet the power to prevent anyone else from touching or
looking at their code?

They developed it.

See my other postings and the reply about division of labour. You might
also read up on Maos Great Leap Forward and north coreas policy of doing
everything themselves.
You're not seriously trying to draw me into to a discusion on
communist history are you? If so, please go ahead, it may be
intersting. I've been reading the Fabian writing of George Bernard
Shaw recently myself.

Right. Mao wanted every village to be self-reliant and do everything on
their own. I think the best published example was that more or less
every village had its own steel factory, resulting in a very low efficiency
and crap steel. If you read about north corea you will sooner or later
stumble on something similar, called "Juche". A fierce desire to be
independent, an inability to recognize you can't be a specialist of
everything and, consequently, a desaster.

By the way, I am _not_ arguing that one must do everything
themeselves, only that one should not get locked into being dependant
on a single provide.

As I'v said, I'm baffled that this is so controversial that you all
expect me to defend my good name merely for saying it.

I'm not sure what this example is supposed to illustrate. The vendor
failed to fix the bug originaly and ony did so under dures,

The point was that contracts work.
It was quite a poorly demonstrated point, as they nearly did and could
well have lost their own customer under the arrangement.

Not "nearly", the legal opinion was correct and therefore the only ones
to worry were the sued ones.

which only
shows how vulnerable you where to begin with,

Why was he vulnerable if he had a contract that required the vendor to work?
Because there is no such requirement,

See my other posting.
As the old joke goes: "if this fire alarm fails, and your house burns
down, we will refund the entire purchase price (not including the
battaries)." OTOH, "if you install this fire alarm, you will pay less insurance on
the house".
if you had the right to
say 'OK, were going to fire you and give someone else the contract'
they would have fixed your bug pronto with no back talk.

No, they wouldn't, because first they would have to understand the code.

Maybe, but in case of open source software they'd say 'Good luck
working into our source code, see you in two years'.
Were do you get this idea? You can contract many companies, large and
small, to support your open source product, the difference being that
you can hire another when when they fail, because you have a right to
the source code, where as you have no recource when the provider has
all the rights.

Like, suse and redhat, each doing their own distributions?
Could you provide a link where IBM actually provides support
for mysql? The only thing I have found is them bragging that MySQL AB
(fully) supports the AIX port, not that IBM supports MySQL.
The only way you can get that kind of support is with a maintance
> contract. With Open Source we'd have had to spend many extra
> man-hours trying to find where the problem was and how to fix it
> without breaking anything else.

Why? You could have the exact same contarct with a vendor supporting
an open source product,

Yes, but then it would cost like any other product, right?
Yes, developing applications costs money, it is this investment I am
advising people to protect by not getting locked into third party
dependencies.

I do get locked into a third party dependency, even if I can change
the third party. I agree, on the plus side, I can change support without
changing code, so who actually owns the code and merges the
fixes from the other guy, provided they don't want to keep them themselves
because they want to keep the customers?
or negotiate access to source for the vendors
product, the only difference being that you then have leverage.

The access to the source means nothing, see above.
It means everthing.

Why? I can't change it.
It means the difference being being the master of
your applications and contracts or being a slave to a third party
vendor. He's my slave because I pay him.
Or
failing that, your application could have been designed to to give you
alternatives,

Right. And the customer throws away years of experience with one db system
and pulls a finished, reliable and maintainable alternative installation out
of the hat.
Maybe not 'out of the hat' but with less expense and retraining that
having to reprogram the entire application which was programmed with
proprietary bindings everwhere instead of properly abstracted code.

Abstraction can make the job easier, you are right here, but then
changing a database is not that hard too, as long as both are relational ones.

Including people who have been trained on it.
In what way is a change from oracle to db2 easier than a change from
postgresql to mysql?
Well, for one, you would never have to change away from the open
source products because of a dispute with the developer.

Yes, I would. Because I'm not going to maintain my own database
distribution.
But in
anycase, my argument is not, and never was, oracle and db2 versus
postgresql or mysql. But rather for abstraction when you do not have
source code, or sometimes then too. If I have abstraction it's even less necessary to mess around with
the db because it's easier to change the db.
But you put yourself in a position were you may have been unable you
support your own customer _AT_ALL_ except for the good graces of your
vendor.

Why? He doesn't support the db. The db vendor does that. All he has to do is > to show that it's othe db's fault, at which point

his customer's maintenance contract with the db vendor kicks in. Normal business practice.

Yes, passing the buck is unfortunalty the normal business practice,
however good firms neither do it or put up with it.

And that is why special libraries, databases or servers exist?
I certainly would
not expect my clients or users to be satisfied when I told them, I'm
sorry the application I provided for you doesn't work, but you will
have to discuss this with Larry Ellison. Nor would I be satisfied
giving such an excuse.

It's different for databases.
A) the customer quite often already has a database and expertise
maintaining it. He has an interest not to have another.
B) the customer may trust Larry ellison, or IBM more than me.
C) the customer may want a database that can do more than I could
implement or maintain, like incremental backups, logical/physical
standby databases or security.
Another case where it's different would, for instance be the OS.
How much linux maintenance do you think you can provide,
compared to redhat or suse? Is this really your corebusiness
or area of expertise?

Greetings!
Volker

[Volker Hetzer]

"Steve" <st**********@y ahoo.com> schrieb im Newsbeitrag news:6f******** *************** ***@posting.goo gle.com...

Daniel Morgan <da******@x.was hington.edu> wrote in message news:<108369595 7.163784@yasure >...

Sarah Tanembaum wrote:

Beside its an opensource and supported by community, what's the fundamental
differences between PostgreSQL and those high-price commercial database (and
some are bloated such as Oracle) from software giant such as Microsoft SQL
Server, Oracle, and Sybase?

Is PostgreSQL reliable enough to be used for high-end commercial
application? Thanks

PostgreSQL is highly overrated and not suitable for any environment
where little things like crash recovery and security are a priority.

What database does Google use?

They offer jobs maintaining a "Linux cluster consisting of more than 10,000 servers".
I doubt that any single database scales that far.

Lots of Greetings!
Volker

[Quirk]

Daniel Morgan <da******@x.was hington.edu> wrote in message news:<108425346 7.342535@yasure >...

Probably. But only right up until it crashed or some cracker tried
to break in. Then it would likely be both as fragile and as transparent
as a sheet of glass.
FUD

"As a cryptography and computer security expert, I have never
understood the current fuss about the open source software movement.
In the cryptography world, we consider open source necessary for good
security; we have for decades. Public security is always more secure
than proprietary security. It's true for cryptographic algorithms,
security protocols, and security source code. For us, open source
isn't just a business model; it's smart engineering practice."
-- Bruce Schneier, Founder and CTO Counterpane Internet Security, Inc.

http://www.schneier.com/crypto-gram-...rceandSecurity

"Microsoft is really good at producing really cool stuff. Security
isn't cool, I want to produce good stuff and customers want dancing
pigs."
-- Carl Ellison, security architect at Microsoft Corp.

http://www.eweek.com/article2/0,1759,1386333,00.asp
Does anyone really believe that if SAP and PeopleSoft could make as
much or more money writing their products to work against these
products they wouldn't?
SAP does:

http://www.mysql.com/news-and-events...e_2003_16.html
Does anyone really believe that CFOs and CIOs,
looking at their budgets, wouldn't be running to these products en-mass?

They are.

http://www.dwheeler.com/oss_fs_why.html

[Quirk]

"Volker Hetzer" <vo***********@ ieee.org> wrote in message news:<c7******* ***@nntp.fujits u-siemens.com>...

What cold comfort that is. I would prefer the right to make my
aplication work without their good graces.

Before you consider suing them I suggest you reiview your contact with
an actual lawyer. So you can understand exactly how painted into a
corner you really are.

Look, we've got about 50 people here dealing with
exactly those questions, telling us what contracts to enter and what not.
When we buy support, we *know* what we are in for and when and
what to sue them for and how to deal with them before we sue them.
Your argument, as usual, is that I should just believe you, not
because you have explained yourself, but just because you *know*.

Wether you have 50 people or 100 people 'around there', the fact
remains that it is very unlikely that your investment can be saved by
a lawsuit, for every 50 you have, Oracle has more. And if you do have
more legal might than Oracle, you are the exception, not the rule.

For most organisations, sueng Oracle, or anyother major corporation is
simply not an option.

My orignal comments still hold true, the right to sue is cold comfort,
the right to pick up your pieces and try somewhere else, keeping your
application in tact as much as possible, is better.

at least, the right to cancel the
contract which hurts them way more

How can you cancel the contract when your entire application is
dependanton there product? Can you afford to throw away your
application too? See my other posting. Compared to changing the application (replacing
it with another), changing the underlying database is easy.
Even easier if you have abstracted your data access with a simple
function, and then used that function throught your application. I
have no idea why you find this so hard to believe.

And for what purposes are you bringing up changing the application?
How is this comparison relevent? I am trying to explain how to protect
your investment in your application; to change it as little as
possible.

You make so little sence I wonder what is motivating you to carry on.

Abstraction of your database access is a good idea. Why are you so
hell bent to dispute this.

than if I go to a postgres developer
and tell him I'm not interested any more. So, unlike open source developers,
they actually have an interest in doing something.

What on earth are you trying to say here? Why is a postgresql
developer any more or less interested in your contarct than one who
pedals proprietary software? A developer who does not earn mony by it is less interested in providing
work than one who does.
Why would anyone provide work for you without earning money? Geez, I
feel like I should be earning a paycheque just for talking to you.

As I've said, their are many profesional developers who provide
support for open source products, or provide source licences for their
own.
Therefore, support contracts make sense.
Of course they do.

They make even more sence if you are not locked in to a single source.
I was talking about the case where I go to the developer and ask him
to do something for free.
Why would anybody do work fo you for free? Are you a charity of some
sort?

Why? What magic powers are possesed by the firm that holds the
copyright? Expcet the power to prevent anyone else from touching or
looking at their code? They developed it.
Not necessarily, they merely own the copyright. And even so, that
still does not mean that somebody else can't modify it, and do so
well, sometimes even better than the original developers.

See my other postings and the reply about division of labour. You might
also read up on Maos Great Leap Forward and north coreas policy of doing
everything themselves.

You're not seriously trying to draw me into to a discusion on
communist history are you? If so, please go ahead, it may be
intersting. I've been reading the Fabian writing of George Bernard
Shaw recently myself. Right. Mao wanted every village to be self-reliant and do everything on
their own. I think the best published example was that more or less
every village had its own steel factory, resulting in a very low efficiency
and crap steel. If you read about north corea you will sooner or later
stumble on something similar, called "Juche". A fierce desire to be
independent, an inability to recognize you can't be a specialist of
everything and, consequently, a desaster.
And the relevance of this is....?
By the way, I am _not_ arguing that one must do everything
themeselves, only that one should not get locked into being dependant
on a single provide.

As I'v said, I'm baffled that this is so controversial that you all
expect me to defend my good name merely for saying it.

> I'm not sure what this example is supposed to illustrate. The vendor
> failed to fix the bug originaly and ony did so under dures, The point was that contracts work.

It was quite a poorly demonstrated point, as they nearly did and could
well have lost their own customer under the arrangement. Not "nearly", the legal opinion was correct and therefore the only ones
to worry were the sued ones.
If it did come to a dispute, they could not have supported there own
application, they where exclusively dependendant on an outside firm.

> which only
> shows how vulnerable you where to begin with, Why was he vulnerable if he had a contract that required the vendor to > > > > work?
Because he had no right to go elsewhere if the vendor failed to
deliver.
As the old joke goes: "if this fire alarm fails, and your house burns
down, we will refund the entire purchase price (not including the
battaries)." OTOH, "if you install this fire alarm, you will pay less insurance on
the house".
Relevence? What insurance is provided in the case here?
Fire insurance you can buy, I have never heard of application
obsoletion insurance.

The original point being, you can not recoup your own investment, just
the purchace price.
> if you had the right to
> say 'OK, were going to fire you and give someone else the contract'
> they would have fixed your bug pronto with no back talk. No, they wouldn't, because first they would have to understand the code.
If they where a credible provider of support and development for this
particular product, they would certainly understand the code.
Maybe, but in case of open source software they'd say 'Good luck
working into our source code, see you in two years'.

Were do you get this idea? You can contract many companies, large and
small, to support your open source product, the difference being that
you can hire another when when they fail, because you have a right to
the source code, where as you have no recource when the provider has
all the rights. Like, suse and redhat, each doing their own distributions?
Huh? No, like a competent development comany providing devlopment
services, exactly like Oracle does, but without trapping you into a
sole source situation.
Could you provide a link where IBM actually provides support
for mysql? The only thing I have found is them bragging that MySQL AB
(fully) supports the AIX port, not that IBM supports MySQL.
Your question is yet another fallacy, since you are responding to a
general statement, that many large companies, including IBM, support
open source applications or provide source licences for there own
applications, but if you really want to hire IBM to support your
MySQL implemtation, you can, I would recomend you try MySQL AB first
though.

IBM Application development and systems integration
http://www-1.ibm.com/services/us/ind...t/bcs/a1000402
Yes, developing applications costs money, it is this investment I am
advising people to protect by not getting locked into third party
dependencies. I do get locked into a third party dependency, even if I can change
the third party.
If you can change it, you are not 'locked in.'
I agree, on the plus side, I can change support without
changing code, so who actually owns the code and merges the
fixes from the other guy, provided they don't want to keep them themselves
because they want to keep the customers?
All these question depend on the case, and have nothing to do with the
topic, if you have a right to the source you are safer that if you do
not, if you have abstracted your access you are safer still. What is
it you can not understand?

This conversation is becoming surreal.

> or negotiate access to source for the vendors
> product, the only difference being that you then have leverage. The access to the source means nothing, see above.

It means everthing. Why? I can't change it.
You have the *right* to use it and have it changed for ever and ever,
not only by the permission of some outside company.
It means the difference being being the master of
your applications and contracts or being a slave to a third party
vendor. He's my slave because I pay him.
No, he can simply ignore you if he decides the relationship is no
longer profitable for him. You can do nothing.
Maybe not 'out of the hat' but with less expense and retraining that
having to reprogram the entire application which was programmed with
proprietary bindings everwhere instead of properly abstracted code. Abstraction can make the job easier, you are right here, but then
changing a database is not that hard too, as long as both are relational
ones.
That's all I'm saying, Abstraction is a good idea. I was giving some
simple, good advice. What are you saying?

Including people who have been trained on it.
In what way is a change from oracle to db2 easier than a change from
postgresql to mysql?

Well, for one, you would never have to change away from the open
source products because of a dispute with the developer.

Yes, I would. Because I'm not going to maintain my own database
distribution.
Nobody asked you to. You have the right to use the product and never
talk to the developer if you like. You don't need to change it to
enjoy the rights that source code gives, that is the right to use the
product for ever, and even have it changed *if you need to*

My advice is to abstract when you have no source code, and perhaps
even then, I have repeated this many times and am not sure what you
are even disputing.
But in
anycase, my argument is not, and never was, oracle and db2 versus
postgresql or mysql. But rather for abstraction when you do not have
source code, or sometimes then too. If I have abstraction it's even less necessary to mess around with
the db because it's easier to change the db.
Yes, that's why I am *recomending* abstraction. Are you just typing
compulsively at this point?
I certainly would
not expect my clients or users to be satisfied when I told them, I'm
sorry the application I provided for you doesn't work, but you will
have to discuss this with Larry Ellison. Nor would I be satisfied
giving such an excuse.

It's different for databases.
A) the customer quite often already has a database and expertise
maintaining it. He has an interest not to have another.
Abstaction means your application can run for different clients with
different databases then. double plus good.

However if your application is tied to one database, then the very
client you are describing is the very client that you will not get if
they use a different database from yours.
B) the customer may trust Larry ellison, or IBM more than me.
But if they only sent there money to Lary because they purchaced your,
unabstracted application, they would be pissed off when it did not
work, and you blamed it on Larry.
C) the customer may want a database that can do more than I could
implement or maintain, like incremental backups, logical/physical
standby databases or security.
Exactly, so how are you going to accomplish this with your
unabstracted application? Do you even remember what side of this
debate you are on?
Another case where it's different would, for instance be the OS.
How much linux maintenance do you think you can provide,
compared to redhat or suse? Is this really your corebusiness
or area of expertise?

Why do I have to? Since I can hire one of a million support providers
for any OS, however for OSes without source, they can't do much when
the problem is with the OS itself. Same with the database.

Again, my argument summerized for the millionth time: If you have no
source Abstract access for sure, and it's also a good idea to abstract
access even if you do. I'm baffled how you've turned this into such a
long conversation.

[Erland Sommarskog]

Dmytri Kleiner (qu***@syntac.n et) writes:

The Application that you wrote I have no reason to doubt is of
sufficient quality to keep your customers satisfied.

Unfortunately you have created unneeded dependencies for them, the
worst of which is not MS SQL, since it is fairly easy to get at data
in MS SQL and archive it or export it in a usefull way, the worst is
that you have tied your customers to a terrible Operating System with
a terrible licence, even Oracle users are not so screwed since at the
very least they have a choice when it comes to OS.
The fact that you may found Windows a terrible operation system is
of course completely irrelvant to the discussion.

If it wasn't clear: we offer our customers a product, and they are not
only tied to the DBMS and operating system, they are just as well tied
to our product. They can still change a competing system, and this
has happened, for instance in conjunctions with mergers. (In which case
it is more an issue of politicis and which company that buys which
that determines which system they go for, than the technical qualities
of the respective systems.) Converting data from one system to another
is of course a major task.

As for the platform, the customers knows what they get when they buy
our system. If they don't accept Windows, they are not likely to go
for us either.
I would say you made it quite clear that your basic message was that
it would be folly to do what I was suggesting,
Yes, it would be a folly to do so out of principle always. Sometimes
it may be necessary, sometimes you are better off tying yourself to
one single platform.
It is not, as I've said, it can be as simple as writing a wrapper
function around your data access.
Yes, if you build your system with all logic in a middle layer. Which
often can result in serious performance problems, because a lot of
data has to travel forth and back over the network. We have a lot of
the business logic in stored procedures, and we have also found that
this works best.
Not as expensive as having the system itself obsoleted by an obsoleted
dependency or the inabilty to get support for a dependency due to a
licencing dispute.

Well, my company has worked this system since 1992, and nothing close
to that has happened yet.

--
Erland Sommarskog, SQL Server MVP, so****@algonet. se

Books Online for SQL Server SP3 at
http://www.microsoft.com/sql/techinf...2000/books.asp