Skip to Main Content

Breadcrumb

Question and Answer

Tom Kyte

Thanks for the question, Dana.

Asked: November 20, 2004 - 6:31 pm UTC

Last updated: July 06, 2005 - 7:51 am UTC

Version: 10.0.1

Viewed 1000+ times

You Asked

Is Tom (the person, not the website) 24x7? When do you do maintenance?


and Tom said...

LOL.

I sleep in whatever timezone I'm in for at least 6 hours (well... i try)... I prefer 8, totally.

the more I travel the more 24x7 i appear to be though. I really liked being in Norway. When I answered email at 6am there, it was 10pm the night before back home. I could actually answer email and not have new mail when I was done :) how cool is that..

I can see the US ad campaign now "come to europe, you can answer your email safely here"....

I did shutdown the site for two weeks recently -- end of september beginning of october... Too much travel at that point. So, way way below 24x7.

Rating

  (38 ratings)

Is this answer out of date? If it is, please let us know via a Comment

Comments

How about help.

Sai, November 20, 2004 - 9:01 pm UTC

Hi Tom,

In one of the threads you said from time to time you do take help from some of your collegues. In other words, do you take help from Oracle product development group, or access internal documentation which is confidential to product development group only. The reason why I am asking this question is it does seem like one need access to such documentation to answer some of the toughest questions on this forum. Is this true, or you only access the documentation we have access to.

Thanks for your time.

Tom Kyte
November 21, 2004 - 8:54 am UTC

give me an example of a question that you felt would necessitate inside knowledge....

I've never looked at source code for the database (well, i did once, trust me, it would not be useful)

I don't dump blocks (well, once i did -- to show when a lob moved out of line..)

I don't use "inside knowledge" really.


I try to show how I came to make the statements I make -- via 'test cases' that you can reproduce (or not -- that is just as important, finding the cases where they do not reproduce, those are the exceptions, the caveats we must know about stuff. Nothing is 100% true or false -- just information we can amass and apply over time)....

I do not believe (i cannot recall) any time where I've used "inside information" about the workings of Oracle -- stuff you would not be able to glean yourself.

The help I ask for manifests itself on this site with an answer that begins:

I've asked Sean Dillon, my XML expert ...
I've asked Clarke Colombo, my Spatial expert...
I've asked Tyler Muth, our resident HTML DB expert...
I've asked Joel Kallman, our resident Text expert...
I've asked <so and so>, our resident <whatever> expert...

so the answer is fully attributed to them -- and they use the same format I do -- proof by example with test case.

lifting the hood

Alberto Dell'Era, November 21, 2004 - 10:52 am UTC

For sure you can understand how a car behaves without knowing anything about what's under the hood, by experimenting with the steering wheel, brakes pedal and so on, and you can become even a F1 driver that way...

But wouldn't it quicker to get the car manufacturer "internal documents", and understand how the engine is built - to get a general idea ?

Out of metaphor, after having read Steve Adam's book, the closest thing to an internal doc i have, my learning speed was greatly increased - because i had a better idea of the engine architecture. The same for Wolfgang Breitling's doc about the CBO, and for a block dump showing the freelists ...

An example of a question that could be easily answered by reading an internal doc is this q of mine:

</code> http://asktom.oracle.com/pls/ask/f?p=4950:61:::::P61_ID:17989405187749

for sure in some internal doc there's a picture or description analogous to

http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:17989406187750#18300957395464 <code>

that imho would be immensely useful.

Have i changed your mind ? (probably no ;)

Tom Kyte
November 21, 2004 - 7:09 pm UTC

but those "internal" documents - any guess as to how they were themselves developed? with "internal" knowledge?


i don't see how an internal's doc would help with the first link at all-- it is a fact that it worked via "method A" in versions prior to 9205 -- with "method B" in 9205 and later. All we can do is measure the impact of the change. Having access to tons of internal "stuff" (like source code) would be like looking for a needle in a haystack.

The second one isn't a call for reading "internal documentation" in as much as it is "hey, how's about the gory details". that is, it would not be an internal "oracle eyes only" document that would be called for.


The question was "Tom, do you access internal documentation which is confidential to product development group only."

Answer is "no, i do not". I do it the same way Steve Adams, Wolfgang Breitling, Jonathan Lewis, Cary Milsap, et.al. do.

Revisiting the car analogy

Bob B, November 21, 2004 - 8:33 pm UTC

If you were trying to drive a car for a race, you'd use the car in race situations and try out different things to find the best way to race the car under that "version". After racing, if the performance was suboptimal, then you'd look at the "internal" documents for the engine, transmission and whatnot and experiment on those to come up with the best performance.

"External" documentation (public published documents) help you find the best/correct way to run the oracle "engine" that you got. "Internal" documentation helps the oracle developers tune and improve the oracle "engine" itself. Of all the questions Tom has answered that might require "internal" documentation, the person has been asked to file a tar with support.

Alberto Dell'Era, November 22, 2004 - 7:26 am UTC

>The second one isn't a call for reading "internal documentation" in as much as
>it is "hey, how's about the gory details". that is, it would not be an
>internal "oracle eyes only" document that would be called for.

But to get to the gory details - wouldn't it be easier to read the docs describing how that feature is implemented, before experimenting ? Not the source code, just the architecture of the various caches (pl/sql cursor cache,session cursor cache) etc. in that example.

I agree that internal docs or undocumented things are not "sorcerer's secrets" that once read, will make us run faster; only, sometimes, imho, a lift of the hood, a reading of the car's blueprints, may be beneficial.

btw, imho the architecture of the various caches *should* be documented externally...

Tom Kyte
November 22, 2004 - 7:50 am UTC

actually, the trend is towards less internals type documentation -- just like cars :)

I own a fairly complex car. A gas/electric hybrid. Beyond "take the car in for servicing by a professional when you need an oil change" -- there really isn't any internal documentation. They spend pages on how to operation the navigation system - and cursory information on how the engine works.

cool car

graeme king, November 22, 2004 - 9:51 am UTC

tom

that is great to hear that you own such a environmentally friendly car.




Hmmmm...when "internals" become "externals"

Connor, November 22, 2004 - 9:57 am UTC

Consider a pet view of mine: ora_kglr7_dependencies

Yes, its an internal view. It appears somehow related to RPC calls and parsing. Normally we wouldn't care, but at a site I work at, its commonly the dominant consumer of logical IO. We've hinted the view to lower the cost, but that's officially unsupported and not the way we like to do things.

Nothing in asktom, our own OakTable, or tar's raised with Oracle support yields a definitive answer about the conditions under which its called.

I'm *not* criticising that fact - its still on the "to do" list for personal R&D, but it does raise the fact that sometimes the "internals" would be very useful to solve an externally visible problem.

Cheers
Connor

Arun Gupta, November 22, 2004 - 10:04 am UTC

Tom,
I fully agree with you when you say that knowledge of internals or looking at source code would not be very fruitful. Most problems do not require ripping apart the source code, just basic principles, as we learnt very recently (and learn everyday). Bugs are easy to spot, they are either documented or there is a marked deviation from documented behaviour or ora-0600, 7445, 3113 etc, for everything else, it requires only logical thinking and asktom.oracle.com !!

However, I am curious about something. I have been reading this thread about index rebuilds. Many experts say this is not required, some say it is required. Similarly, there is a big thread on difference between select count(*) vs. select count(1).

In such cases and in other cases of other Oracle myths, wouldn't a comment on the internals help settle controversy faster?

Thanks

Tom Kyte
November 22, 2004 - 3:15 pm UTC

No -- it would not.

Take index rebuilds. I believe you've taken some quotes out of context. I'll paraphrase:

there is one camp that says all indexes rebuilt on a schedule (recurring, over and over, every one or most of them) is a waste of time, energy, resources and could lead to decreased performance as often as it leads to increased or no change in performance.

there is another that says "Index rebuilds are low risk", "Index rebuilds are unobtrusive", "Index rebuilds are cheap"

I think this:
</code> http://www.jlcomp.demon.co.uk/circular.html <code>

points out some of the flaws in "camp 2" there. but -- how would knowing the nitty gritty details of the indexing code in Oracle help you come to the answer/conclusion? see in that case, the totality of the overwhelming amount of detail would make you "lose sight of the forest for the trees".

I'd still be back to setting up simulations saying "this is when it breaks down and how", "in this use case -- a rebuild should be considered perhaps", "in this case -- a rebuild is an utter waste of time - even it was was FREE (which it aint') it would still cost too much", "in this case, rebuilding is a truly BAD idea".

for the index rebuild, the simulations of the various use cases are what we need to know - not the internals.


Same for count(*) vs count(1)

the statement is generally "count(1) is faster or cheaper"

the answer is always "no it isn't, see, here is the example in many different forms showing they are the same. Tell you what-- present the SINGLE test case that shows a use pattern whereby count(1) is superior to count(*)"

I've yet to recieve that use case ;)

A reader, November 22, 2004 - 11:13 am UTC

Thats the beauty of asktom, the answer to almost every question is already here.

Your rebuild index question.
</code> http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:2290062993260

and you count(*) and count(1) question
http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:1156159920245 <code>

Alberto Dell'Era, November 22, 2004 - 4:09 pm UTC

> "take the car in for servicing by a professional .."

But i am/wanna be that professional ... and surely i need more than the car operating instructions to perform my work well, unless i want to play
Voodoo [ hint to c.d.o.s ;) ] with the car ...

Well, if i had to suggest an improvement to "Expert 1on1", that would be to write more about the memory architecture of Oracle (what's a library cache handle, what's a pin, a get, etc). The level of detail doesn't need to be too deep, e.g. just a bit more detailed than this following answer by JPL:

"On a typical 'get', the session has to acquire a 'pin' structure on the buffer header before reading the buffer. (The sequence is: get latch, search chain, pin buffer header, drop latch, read buffer, get latch, unpin buffer header, drop latch)".

</code> http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:1229436447262#28741138524830 <code>

Another example may be the pl/sql vs session cursor cache architecture ...

Information like that *immensely* helped me in many many ways (while designing, not only while "tuning"), so i feel that it deserves a tad more coverage ... it's architecture after all, imho no less important than the Processes or redo or undo architecture.

Just my 2cents - i don't wanna "push" you, just suggest and discuss.

Tom Kyte
November 22, 2004 - 5:27 pm UTC

any my belief is that 99.999% of the time that is "TMI" (too much information) and at such a level of detail that, well, it would not be too big of a seller.

I'm going for the low hanging fruit -- I wish I could get everyone just to read the *concepts* manual. We've got miles and miles to go before that level of detail will be mainstreamed by me anyway.

Alberto Dell'Era, November 22, 2004 - 5:38 pm UTC

Thanks for your feedback - appreciate it, even if we don't share the same view :)

You say...

A reader, November 24, 2004 - 12:26 am UTC

"give me an example of a question that you felt would necessitate inside knowledge...."

Like this..

</code> http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:3561623983484, <code>

for example ?

_COMPLEX_VIEW_MERGING .. mmmm...


Tom Kyte
November 24, 2004 - 7:07 am UTC

hmmm indeed:

...
In 8.0.4 - 8.0.6 -- you would set "complex_view_merging" = true (init.ora or
session level) parameter to get this behavior (I setup a small test case to see
this). In 8i, this parameter is obsoleted but _COMPLEX_VIEW_MERGING exists (and
is set for all Oracle application instances so it is widely used. Please
contact support for any issues you might need to be aware of when using this
parameter as it is undocumented).
.........


Lets see:

o in 8.0 it was a fully documented init.ora parameter.
o in 8i it was documented as deprecated (but still existed as an _)
o Oracle Applications publicly documented "you shall set this"
o metalink discusses it



Nope, not "inside knowledge", only stuff you could figure out yourself.


Try again :)


internal knowledge required for ...

Alberto Dell'Era, November 24, 2004 - 8:17 am UTC

</code> http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:11504247549852#20191884590175 <code>

"... and some consultation just to make sure ..."

;)

Tom Kyte
November 24, 2004 - 8:58 am UTC

nope, not a single bit of internal knowledge needed there -- the example speaks for itself?


My point here is "No, you do not need access to source code, no, you do not need access to the people that wrote the code, no, you do not need to dump blocks (often :), no, you don't need internal 'eyes only' documentation"

I don't anyway -- people assume that I use it all of the time, when quite the converse is true -- I've never used it. I won't use it. I use the same techniques and tools *anyone* out there has access to.

A reader, November 24, 2004 - 10:18 am UTC

If Oracle internal documentation is indeed necessary, how people like Jonathan Lewis, Carry Milsap and Steve Adams do what they are able to do? None of them have access to Oracle source code or internal docs. (BTW, Steve's great site is kind of inactive for over an year now, a standing testimony that there are no takers for the internal info anymore because you don't need it that much)

A reader, November 24, 2004 - 10:18 am UTC

If Oracle internal documentation is indeed necessary, how people like Jonathan Lewis, Carry Milsap and Steve Adams do what they are able to do? None of them have access to Oracle source code or internal docs. (BTW, Steve's great site is kind of inactive for over an year now, a standing testimony that there are no takers for the internal info anymore because you don't need it that much)

A gas/electric hybrid

Vladimir Andreev, November 24, 2004 - 12:18 pm UTC

Now I understand your obsession with counting red lights between your home and work -- you probably charge your battery each time you brake :-)

Imagine -- each time we spin on a latch, we do useful stuff...

Wouldn't that be wonderful!

Tom Kyte
November 24, 2004 - 12:27 pm UTC

LOL, if the energy spent spinning on a latch could be captured and turned into something useful -- applications that didn't use bind variables would actually be power plants!

Alberto Dell'Era, November 24, 2004 - 1:28 pm UTC

Hey, i've found your car :)

</code> http://www.cruise-south.co.uk/tom/tom.htm <code>

Tom Kyte
November 24, 2004 - 1:57 pm UTC

ROTFL - don't know which is racier, the cars or ...


actual car is </code> http://www.toyota.com/prius/ <code>

24X365?

Peter, November 25, 2004 - 11:08 am UTC

I see that you have been responding to queries - don't you have a turkey to carve - or is my UK diary wrong and it is not Thanksgiving?

;-)



Tom Kyte
November 25, 2004 - 11:52 am UTC

Oh, it is -- turkey in the oven, 2 hours to go... Best smelling day of the year.

personal web site

Prasad, November 25, 2004 - 6:18 pm UTC

Tom,

Do you have a personal web site?

Prasad

Tom Kyte
November 25, 2004 - 8:17 pm UTC

asktom.oracle.com

it is very personal to me :)

Navaigation = best thing ever put in car

robert, November 25, 2004 - 8:32 pm UTC

somehow I pictured Tom as a BMW/Lexus guy - sport sedan optimized for perfomance. A hybrid like Prius is optimized for fuel economy above all else.

As for Nav..personally I will never ever get another car without it. Totally worth every penny. Love it !

I'm going to test drive Prius to check out to quench my curiosity.


Tom Kyte
November 25, 2004 - 8:44 pm UTC

I had a (used when I bought it) 94 BWM 5 series. (still have it, it is our extra car for emergencies/when friends need it/whatever)....

Lots and lots of miles on it. Loved it.

However, the opportunity to double my miles/gallon....

with more interior room for all....

at about 1/3 the price (i looked at new 5 series -- wow, i just cannot pay that much for a *car*. computer equipment, something useful -- sure, but a *car* -- no)

that and the kids think this car is the coolest car ever :)

It might not beat out a 2005 BMW series for luxury -- but, I really dig paying 10/15$ USD for a tank of gas, and riding the HOV lanes when I want (i live in virginia, i can ride the car pool lanes :)

The fact it emits the least emissions possible makes me feel rather good too. We are here for a brief time, our emissions are forever.




personal web site and 24x7

Prasad, November 25, 2004 - 11:05 pm UTC

I knew you may say point this site. You must be 24x7. What's the time when you answered by prev. question? around 4.00 AM?? Please don't get sick... else thousands, if not millions, of others die of heart attack!

Tom Kyte
November 26, 2004 - 9:02 am UTC

i'm on the east coast of the US, as I recall -- it was about 8pm I did that one, after everyone left from the thanksgiving dinner

I haven't read your books, but..

A reader, November 25, 2004 - 11:08 pm UTC

Sorry Tom, I haven't read your books as of yet (waiting for them), but I do have a personal question for you. Do you have an university degree ? and If so, which is it ?

You are a genius!

Tom Kyte
November 26, 2004 - 9:22 am UTC

I'm most decidely not a genius.

Yes, I have the standard 4 year degree from the University of Pittsburg -- in mathematics. I did not really touch a computer until *after* graduation. Our interface to the VAX the university had back then was a teletype believe it or not ;) try editing files on a teletype -- no monitors......

I saw an advertisement for "programmer wanted, no experience required". I met all of the criteria (I had absolutely no experience). Not much else you can do with a math degree -- except get more degree's ;)



Top notch car? No way!

Gj, November 26, 2004 - 3:34 am UTC

You think Mr Kyte would have a top range car? Have you learned nothing from the forums? Buy something simple and easy, then work it hard to get the best out of it!

"a simple car!?"

Flado, November 26, 2004 - 5:36 am UTC

No way. If anything, these cars are very complex. That's the biggest problem with them, actually, which makes them relatively expensive. But I'm glad to hear that something is finally being done to reduce emissions on the other side of the pool :-)

Documentation only??

Saibal Ghosh, November 28, 2004 - 11:51 am UTC

Tom,
In many, perhaps most cases and in many threads you urge us to read the documentations. Tell me how do you remember all the stuff you read there, and besides, don't you at times get overwhelmed by the sheer volume of information out there? Since you answer an incredible range of questions on Oracle from all and sundry, I guess you remember quite a lot of what you read. But, my qustion is how and when do you develop all those test cases and benchmarks?

Tom Kyte
November 28, 2004 - 12:31 pm UTC

I have a cursory knowledge of the documentation -- I've at least browsed it.

Now, someone asks a question, i say "you know, i read that somewhere".... goto otn, search for it, find it.

I have not memorized the documentation, I have browsed most all of it at some point or another. Knowing the feature exists, the information is out there is what is key. Not "knowing the information", but rather "knowing the information exists"

did you know, I myself refer to "Expert One on One Oracle" and "Effective Oracle By Design" frequently -- it amuses people when they walk into my office and I have my own book out. I haven't memorized it all, I still look things up! That is why I push people to the doc's so hard and so frequently. They actually have the material in them.


The test cases come naturally -- it is just the application of the scientific method we all learned (hopefully) in grade school/high school.... You have a hypothesis, you lay out a set of assumptions (conditions you are relying on -- they are as important as the test sometimes because of the assumptions are not met -- the test might not be valid, meaning nothing is ALWAYS true in all cases..).


A reader, November 28, 2004 - 1:24 pm UTC

Your response is excellent. Thanks Tom as always for all the support you provide to Oracle community and guiding in the right direction.



Effective Stress Management

Alberto Dell'Era, November 28, 2004 - 1:56 pm UTC

If you don't mind - always plugged to asktom, flying here and there, writing books, etc - how do you manage your stress level ?

It's not a curiosity really - actually i'm hoping that your recipes may apply to me too ...

Tom Kyte
November 28, 2004 - 2:17 pm UTC

Martini's mostly.

as for the "recipe" :)

</code> http://www.webtender.com/db/drink/3069 <code>
(minus out the vermouth, double the cardinality of olives)

Alberto Dell'Era, November 28, 2004 - 2:55 pm UTC

<quote>
</code> http://www.webtender.com/db/drink/3069 <code>
</quote>

ROTFL (and ROTFL again :)

I'm going to put it in production immediately, without Questioning Authority for this. If you don't hear from me anymore, it's probably because i've reversed the Martini/olive ratio :)

don't take a genius to RTFM

robert, November 28, 2004 - 3:44 pm UTC

>> You are a genius!

geniuses go to CarnegieMellon :p



Too much change too fast?

Kal, November 29, 2004 - 7:26 am UTC

Dear Tom

Fisrt of all let me say i just recently found about your site and i regret that i didnt find it sooner. My question is with oracle changing so quickly and sending out so many changes from 8i to 9i to 10g dont you think that there is too many things happening at once? Im a database administrator myself and ive been one for over 4 years or so. I have done the tests for 8, 8i, 9i and soon 10g but its hard for me to choose an area of oracle to concentrate on. There are just SOO MANY FIELDS??? Please i need some advice as to what to concentrate on i know this might seem like a vague question but i need someone to advise me on what to concentrate on and what not to.

Thanks
Kal

Tom Kyte
November 29, 2004 - 8:28 am UTC

well, why concentrate on 'tests' when you can concentrate on 'experience'?



A reader, November 29, 2004 - 8:51 am UTC

Shaken and not stirred I presume???

And I took you for a single malt Scotch guy!


Experience???

Kal, November 29, 2004 - 9:38 am UTC

Hey Tom

Im doing experience for sure but then comes the question - Shall i concentrate more on learning database internals? or IAS operations? those are my two favourite fields and im really good in programming so that complicates things for me?
what do you think?
thanks
kal

Tom Kyte
November 29, 2004 - 2:59 pm UTC

which do you like better and which appears to have more opportunity for you.

Me, I'm a database guy.... It makes me happy.

Prius is fly by wire

jim, November 29, 2004 - 1:43 pm UTC

Prius's steering is fly by wire. More and more cars are going to it. The new BMW's accelerator is fly by wire also. It makes the car easier to put together and lighter. Heck if it works so well in planes why not cars.

details

Dean, November 29, 2004 - 6:29 pm UTC

The steering is rack-and-pinion. The transmission is by-wire. A by-wire steering system would be far too complicated (read: expensive) to put in a car.

</code> http://www.toyota.com/vehicles/2005/prius/specs.html <code>


I see Tom as my Role Model...

Sanjay Jha, December 11, 2004 - 4:37 pm UTC

Tom,
Let me first thank you for not only Oracle Teachings but a new approach to look at my profession. Before asktom.oracle.com I used to be “do the stuff, document it, read all kinds of books” (I have more than 100 books mostly on RDBMS and Oracle in particular), but never felt very confident nor enthusiastic as it was becoming very monotonous and stereotype. But when I started reading more and more the answers you provide at your forum, I got a new approach and a new way to look at my job – “test and benchmark and explore all the features (not necessarily new but existing ones too)”. I am applying the same approach in other RDBMS, I am supporting (UDB) and finding it very useful.
One thing I have noticed is that Tom's approach will eventually make many Tom Kytes (or already has), no I do not mean VP of Oracle Corp., but a true IT professional who "first gets convinced himself by testing the stuff" rather than just believing the things as it is stated in books/manuals or documents.
I personally have learned a lot and still am learning from Tom's site. I have asked some questions (<20 may be), and each time I have noticed amazing response time, that is what may be has made Tom very popular amongst Oracle Users, beside his distinct style of answering. I must admit that I have in past not searched this forum very well before asking something. And that is one reason I do not ask questions now-a-days (more I go deeper in this site, I feel more and more strongly as how little I know about Oracle, and that motivates me to read more and more!)
Tom has mentioned many a times in this forum as well in his book on how he learnt Oracle and still is learning, wonder why people are so curious to know as whether it is Tom alone or someone else behind this or he has access to internal documents and what not! Tell you what, if one does not have an approach in life be it professional or personal, no matter how much resources one has at his/her disposal, one would fail to deliver the way Tom does!
Tom is doing a wonderful job and he always encourages all of us to read the same sources which he found useful! I am reading Oracle's Concept guide and Performance Tuning guide(which were lying in my office since time eternal and no one had ever read!), beside his two books and can't tell you how much I am learning each day!
Thanks a million Tom for all these!


Car analogy doesn't quite fly

Tom P, December 21, 2004 - 2:58 pm UTC

Besides a developer/dba, my hobby is automotive work- so I don't quite buy into the "car analogy" as regards internal documentation/source code/etc to solve an Oracle question.

Here's why: We're operating the Oracle database engine, just like we operate a car! The key word here is 'Operating'.

We're not fixing the Oracle database engine- if it actually breaks, that's Metalink/Support's job, that's how we get patches. We're not hopping it up- that's Oracle's programmers' job, and that's how we get from 7.3 to 10G. And we're not rebuilding it or restoring it or giving it a new coat of paint.

We're doing all those things (fixing, improving, etc) to our Applications and our Data, not the database engine!

However, with the car analogy, people like me are fixing the car, improving the car, rebuilding the car, restoring the car, and painting the car. The two scenarios don't match.

Going back to the "too much information doesn't help"- you might put 87-octane gasoline in your car, and wonder why it lacks power and makes funny noises when going up steep hills. However, your owner's manual has recommended 93 octane, and you're only using 87 octane. The tech book for the motor will go into compression ratio and timing and the computer chip's A/F tables and the knock sensor and cam grinds and etc, but that's TooMuchInformation. Your friend or a website or a mechanic tells you that the car needs 93 octane, so you put in 93 octane at the next gas station, and ZIP, your car is running better.

Even further, lets apply this idea of "internal documents" to regular programming. As I said, I'm a developer. I use Microsoft's langauge suite, and have been migrating to C#.NET. (Since I use Oracle for the database, it makes for a fun time at seminars; I'm either alone in a room full of Java people or SQL Server people!) I don't see how knowing the machine code for the Intel processor would help me write a better program using C#.NET. It might help me write a better assembly language program, but that's not what I want OR need.

Stick with the big picture!

Tim Hall, January 05, 2005 - 8:42 am UTC

If you simply must have internal information use an open source database.

I've been using Oracle for 10 years and I still feel like a beginner at times. It's hard enough keeping the big picture in your head, let alone trying to master all the internals. Added to that, things have a habit of changing every 18 months with each release.

I have to go with Tom's statements. Everytime you have an issue build a test case. It's the only way to be sure, and it covers your ass when people ask you why you did things.

Also, when you think you've found something you don't understand abstract it and plug in your own solution in such a way that you can flip it back once you've solved the problem (or your lack of understanding). Steve F. talks about this sort of thing from a code perspective here:

</code> https://asktom.oracle.com/Misc/oramag/on-fetching-storing-and-indexing.html <code>

Cheers

Tim...

which db-vendor use for 24x7 ?

Joachim Seekopp, July 06, 2005 - 3:15 am UTC

Hallo Tom,

just read
<QUOTE src=asktom.oracle.com>
NOTE: we'll be down for 4 hours on July 6th, 2005 starting at 6pm US/Eastern time (GMT-4)
<QUOTE>

can I deduce from that, that oracle isn't appropriate software
for 24x7-production enviroenment ?

which database-vendor would you recommend for such a requirement ?

Tom Kyte
July 06, 2005 - 7:51 am UTC

that is a pretty silly statement.

You want 24x7, you can get 24x7, it takes work. it takes money. it takes effort. it takes time.

This site, it is not worth my time, energy, effort or money to do so.

So, I don't.

No questions on asktom.oracle.com in the last time

A reader, October 06, 2014 - 9:47 am UTC

Hi,

I noticed no activities on asktom.oracle.com in the last 5 months.
What is the reason for this?