Monday, January 16, 2006

A new question/answer site...

A new Q&A site just came online. Eddie Awad asked me for feedback on this new site of his last week. Looks like an interesting idea – immediately I saw a question/answer that taught me something new. A method to put a table into “read only” mode in a way. Neat idea there using the check constraint.

The Q&A site could be interesting in that people can post not only the question but the answer. Sort of a way to say “hey, I had a neat idea and would like to share it”. Sort of like this recent blog entry of mine was about.


Anonymous Sokrates said....

great stuff, oraqa !

Mon Jan 16, 11:10:00 AM EST  

Blogger Howard J. Rogers said....

The Q&A site could be interesting in that people can post not only the question but the answer.

You mean, a bit like a forum?

Only, instead of using forum software that is intended precisely to facilitate this sort of question-answer exchange, the site uses Blogging software, which isn't.

There are already a thousand and one Q&A sites (Jonathan's community-based FAQ springs to mind as one of the best), and the Oracle community could do without that being diluted by the arrival of yet another. But when the new arrival is implemented with a bad choice of software in the first place, that's doubly a reason why I shall be giving the new site a very, very wide berth.

Mon Jan 16, 01:25:00 PM EST  

Blogger Thomas Kyte said....

I think it is substantially different from a forum.

Forums tend to have somewhat long and winding paths through them. The Q&A's get bogged down in meandering discussion.

I think this has a neat chance at being a "nugget of something new I learned today" type of site. I think the blog format personally is somewhat ideal for that.

I pick up this and that and the other thing from these blogs. I get a laugh from the oracle-wtf. I get bits of information. From this, I might get something once a week or once a month. But I get something.

But - to each their own. If they didn't have something on the first page that made me go "doh, why didn't I think of that", I probably wouldn't have mentioned it.

Mon Jan 16, 01:33:00 PM EST  

Blogger David Aldridge said....

I see Howard's point, but the problem with issues on forums is that they are very specific -- "I have this set of tables, my query isn't optimizing ..." etc.. The Q&A's would hopefully be much more oriented to general advice than to a specific set of circumstances.

I wonder if it is within the bounds of what passes for good internet etiquette (I'm not going to use the "netiquette" word -- except there) to link to blog articles or whitepapers elsewhere to answer questions: for example "How do I install RAC on Linux" etc?

Mon Jan 16, 02:35:00 PM EST  

Blogger Thomas Kyte said....

I wonder if it is within the bounds of what passes for good internet etiquette

I think that is the third W in WWW.

Mon Jan 16, 02:43:00 PM EST  

Blogger David Aldridge said....

>> I think that is the third W in WWW

In that case I'm going to do it, and if anyone complains I'll tell 'em you said it was OK :D

Mon Jan 16, 03:10:00 PM EST  

Blogger Howard J. Rogers said....

Forums tend to have somewhat long and winding paths through them. The Q&A's get bogged down in meandering discussion.

Well, that's better than misleading people with one-line punchy answers that are just plain wrong!

The site you mention has a question, for example, about how you make a table read-only. "Add a check constraint, disable validate" is the one-line answer.

Whereas, in fact, there is no way to make a table read-only. Alter table X enable constraint blah takes care of the published answer! Sometimes a "meandering discussion" is *needed* to point out that what's being asked (or, worse, what's being proposed as a solution) doesn't actually address someone's real security or business concerns, and that a re-think is required.

Similarly, the answer to 'how do I rename a column' is a one-line statement of the 'alter table' syntax... without a mention of the fact that it's 9i and above only, nor a passing nod to a discussion of whether it's wise to go around renaming columns, and what the consequences on views, materialised views, procedures etc might or might not be.

"Nuggets" I can live with. Emaciated snippets of half-truths and missed points, I can't.

So, yeah: the Blog format is ideal for publishing a neat piece of 'something I learned today', and I use mine to that effect all the time. But here we have a site where *pronouncements* of knowledge -for which a Blog is indeed perfect- need correcting, expanding and elucidating ...and that sort of *dialogue* is not what blogs are good at.

Mon Jan 16, 04:52:00 PM EST  

Blogger Thomas Kyte said....

Howard - we'll have to agree to disagree on this one.

I thought the read only table nugget was brilliant. Really wish I thought of it myself. I've been asked that very question many times before - the solution supplied is pretty simple, elegant and achieves the intended goal. True there is no way to "make a table read only", but this sure makes that table read only doesn't it.

Mon Jan 16, 04:55:00 PM EST  

Anonymous Gabe said....

... but this sure makes that table read only doesn't it.

If sqlldr in direct mode is not a concern ...

Mon Jan 16, 05:03:00 PM EST  

Blogger Thomas Kyte said....

Gabe very smartly said....

Indeed, good caveat. It will leave the constraint in a "not validated" mode as well - meaning a direct path load will effectively take this away.

I'll make a note of that over there.

Mon Jan 16, 05:19:00 PM EST  

Blogger Howard J. Rogers said....

Sorry Tom. You can't be a little bit pregnant, and a table can't be a little bit read-only. Either it is read-only or it isn't. The proposed solution makes the table *appear* to be read-only, provided you don't do direct loads as was mentioned, and provided you don't have someone enabling constraints behind your back... which means that it functionally and effectively is NOT read only, and the answer is therefore 100% wrong.

And if someone were to think that answer locks down their data for all time and for all security purposes, they would be mistaken. Not a little bit mistaken, but totally and profoundly in error.

There is only one way to make a table truly read-only, and that's to store it on a physically read-only medium.

Now, let's not get bogged down, even so, in a specific question. The rather larger point is that this has neatly demonstrated my original argument: that more needs to be said about topics than the "smart answer in a nutshell" blog-style format allows for.

The site was supposed to be "Questions and Answers", not "Questions and partial answers that mislead and need Tom Kyte to clarify rather important caveats".

And that is why those sorts of thing belong on a forum, which is intended to discuss, argue, caveat, and illuminate.

Mon Jan 16, 05:57:00 PM EST  

Blogger Alberto Dell'Era said....

The best format would be a wiki In My Humble Opinion; take a look at

concise, complete, informative ...

Then click on the "discussion" tab near the top of the page to see how they get there - with discussion (peer review, etc)!


Example: Gabe may have added his observation directly on the page, *integrating* it with the rest of the material, perhaps after having discussed it on the "discussion" page of course.

Mon Jan 16, 06:11:00 PM EST  

Blogger Connor McDonald said....

SQL> drop table T;

Table dropped.

SQL> create table T ( x number ) ;

Table created.

SQL> alter table T add constraint CK1 check ( 1=1) disable validate;

Table altered.

SQL> alter table T disable table lock;

Table altered.

ora1dc0n0s:> sqlldr userid=pd71986 control=x.ctl direct=true

SQL*Loader: Release - Production on Tue Jan 17 10:08:58 2006

Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.

SQL*Loader-965: Error -1 disabling constraint CK1 on table T

If you were really paranoid, you could also create a db level trigger to capture 'alter' or 'drop' commands on the table as well.

Mon Jan 16, 09:11:00 PM EST  

Blogger David Aldridge said....

Maybe the emphasis should be on the plural of Answers then -- there are many ways to skin a cat or make a table read only, and multiple answers to almost any question -- there's no reason why they can't be added in.

Mon Jan 16, 10:27:00 PM EST  

Blogger Howard J. Rogers said....

Unfortunately, the plurality of answers don't display equally. That is the fundamental problem with blog-type software: the initial "pronouncement" is stated as the truth, and corrections to it are tucked away under a "comments" link which may or may not ever get clicked.

My point about forum software being more suitable for this type of dialog is precisely that no-one takes an initial post in a forum to be anything other than a starting point and an invitation to discuss. There is no hierarchy of answers, except a purely temporal one. But blog software -the site we're talking about, to be specific- presents entirely wrong answers on the front page, and relegates the corrections to a subsidiary page.

It's the wrong approach. It's misleading. It's erroneous, and it shouldn't be recommended, to be frank.

Mon Jan 16, 11:48:00 PM EST  

Blogger Howard J. Rogers said....

As for a Wiki...

And what do we do when Certain World Leading Oracle Experts edit the wiki to argue complete tosh about clustering factors being an indicator of when to rebuild an index?

It is a fallacy to suppose that all contributions regarding Oracle *facts* are equally valid, but a wiki supposes that they are. In the same vein, an Oracle wiki would suppose that the weight of overwhelming Oracle community opinion is "right", but if so, we'd have been stuck with a Wiki propounding Buffer Cache Hit Ratios for years before finally getting sorted out.

A wiki format would certainly be better than a blog format, but neither beats (IMHO) a forum format. And even that can't beat reasoned argument in thoughtful articles for elucidating real fact instead of quick fixes.

Mon Jan 16, 11:59:00 PM EST  

Anonymous John Scott said....

I think a Wiki can work extremely well and as Tom points out it works in a different way than a traditional forum type thread, in that a Wiki or Q&A can be far more to the point rather than having to read through umpteen messages to get to the salient points (although the corollary to that is that there may also be interesting ideas in the followups).

We setup an unOfficial HTMLDB Wiki at (excuse the advertising)
and by and large the response has been very good. Although as with most online community resources the site is only as good as its content (i.e. you need to inspire people to want to contribute)

Tue Jan 17, 03:19:00 AM EST  

Blogger Howard J. Rogers said....

As I said, I'll take a Wiki over Blogstyle any day, but that doesn't mean it's the right format even so.

But the fundamental problem is anyone ever thinking that Oracle can be reduced down to a set of simple 'in a nutshell' truths.

It was always something of a cliche on cdos that the answer was always "It depends" -but the point of cliches is that they're true. It so often depends on the business requirement that it is precisely the "meandering" discussions of what the business requirement is, or should be, that brings out the salient points people need to be aware of.

There is, of course, a class of question governing basic technical facts and competency that can (probably) be dealt with without worrying about a business context: what is an undo segment, why is there a temporary tablespace, how are extents allocated and so on. Which is why it is doubly depressing to see such a basic one as "how do I make a table read only" incorrectly answered!

In similar vein, I notice the one about "How do I rename a column" doesn't mention the (potentially very expensive) implications of doing so, like invalidating views and the minor annoyance of now possessing materialized views that can no longer refresh themselves. Any answer that purports to deal with the pure technical basics, then, needs to be accurate at a minimum, and should at least make an attempt to acknowledge the fact that actions have consequences!

That's the real issue, I think; and the format of the site is almost an irrelevance by comparison.

Tue Jan 17, 04:07:00 AM EST  

Blogger Alberto Dell'Era said....

And what do we do when Certain World Leading Oracle Experts edit the wiki to argue

A wiki format doesn't mean that the wiki cannot be moderated.

Eg: if someone wants to argue or add, it could post it in the "discussion" page and, if the moderator(s) approves, she can then propagate the contribution on the page.

A wiki can have an "Owner", that sets up the fundamental values of the community (eg adherence to the Scientific Method).

Tue Jan 17, 05:11:00 AM EST  

Anonymous John Scott said....

I completely agree that the vast majority of Oracle functionality can't be reduced down to a series of "bulletpoints".

However the reason the HTMLDB Wiki was started was because the same
questions were appearing again and again on the HTMLDB forum (e.g. "whats the different between v('APP_USER') and &APP_USER."), admittedly people still post that style of question in the forum, but it's good to be able to provide a direct link to an answer on the Wiki (which can be modified and expanded upon by users).

I'm a long time user of Google Groups/DejaNews, infact as a developer I'd probably only be 1% as effective if Google groups disappeared overnight, however I believe it's always good to have different types of reference sources to draw on.

Tue Jan 17, 05:14:00 AM EST  

Anonymous Anonymous said....

Experience rarely comes from other people's mistakes. If we want the oracle community to grow and to grow smarter, we first have to get them out of the "I Suck phase" of Oracle development ("I Suck phase" borrowed from Kathy Sierra's Create Passionate Users blog).

I think its rare that anybody for any skill has ever learned much from reading. Sure, we can learn many "facts" that way, but we rarely get the "experience" that is crucial to the high end of the skill spectrum.

For example: I've practiced martial arts for over 13 years and taught for 6 years. When someone new comes in, I could give them tons of pointers with explanations about every aspect of their movement doing any move. Guess what happens if I do that? Except for a glazed stare, nothing.

What do I do? I tell them to lift their knee and kick with their shin. If they do that in front of me with a modicum of control (i.e. they're not falling down), they get to kick a pad a few times to try it out. Guess who teaches most? Not me, its the pad. They'll only listen to any advice I give if they feel they've "got" the new tools under their belt.

How does this translate to Oracle, forums, Q&A, asktom, etc? The web information is the "master", Oracle is the "pad", and the web surfer is the "student". If the student wants to learn, they're going to seek out the information. Once they have enough information, they'll go and kick the pad with the new information to see how it works.

Yep, all that great advice you gave them won't really sink in until they try it out in Oracle and experience some of the same successes and failures.

As long as Edie moderates the Q&A to ensure that completely bad/wrong answers don't show up (e.g. "To rename a column, log in as sys and change this sys table ..."), his contribution will be a good tool for learning Oracle.

Tue Jan 17, 09:57:00 AM EST  

Blogger Joel Garry said....

And then there is the simple issue of having Yet Another Place To Look For Things.


Tue Jan 17, 10:05:00 AM EST  

Blogger Thomas Kyte said....

And then there is the simple issue of having Yet Another Place To Look For Things.

same could have been said about a site near and dear to me in the year 2000.

It'll either flourish and prosper or whither and die.

I'll continue to let new seeds get planted, help them germinate - and then see what sticks.

Tue Jan 17, 10:15:00 AM EST  

Blogger Joel Garry said....

same could have been said about a site near and dear to me in the year 2000.

Perhaps that site allows driving a dialectic towards a canon, but the blogging paradigm doesn't.

(Oddly enough, canon [from] "is derived from a Hebrew and Greek word denoting a reed or cane. Hence it means something straight, or something to keep straight; and hence also a
rule, or something ruled or measured." So creating a canon via a roundabout dialectic is a funny concept.)


Tue Jan 17, 01:46:00 PM EST  

Blogger David Aldridge said....

>> Perhaps that site allows driving a dialectic towards a canon, but the blogging paradigm doesn't.

Well, I think that's stopped the discussion :D

Tue Jan 17, 05:41:00 PM EST  

Blogger Howard J. Rogers said....

As long as Edie moderates the Q&A to ensure that completely bad/wrong answers don't show up

Define "completely bad". Actually, don't bother... because in doing so, you miss the point. It's the paradigm that's wrong, not the quality of the answers or moderation.

I think its rare that anybody for any skill has ever learned much from reading

OK Tim. Let's just chuck 1000 years of scholarship and experience of scholarship out the window, shall we? Let's just write that if your clustering factor is bad, you can fix it by rebuilding your indexes, shall we? Or that you can make a database perform 400% better by increasing the buffer cache hit ratio from 95% to 98%. Why not? I mean, if you can't learn anything *good* by reading, you surely can't learn anything *bad*, either.

Or perhaps we should better admit that written sources of information get used, have influence, and that they'd therefore better be good ones. Ones that give the right answers more prominence than the wrong ones. Ones that don't turn a complex problem into a stupid one-liner.

As for helping things germinate: weeds germinate too, and they can be a bugger to kill off, and have a tendancy not to whither and die.

Tue Jan 17, 07:01:00 PM EST  

Anonymous Anonymous said....

Somebody on this thread needs to switch to decaf...:-)

Tue Jan 17, 11:16:00 PM EST  

Blogger Howard J. Rogers said....

Ah, the first ad hominem of the debate! I wondered who'd be first.

Wed Jan 18, 12:08:00 AM EST  

Anonymous Richard said....

Another possible method of converting a table (called bowie_ro) to read only like status:

SQL> create table bowie_ro (id number, name varchar2(20));

Table created.

SQL> alter table bowie_ro add primary key (id);

Table altered.

SQL> insert into bowie_ro values (1, 'Bowie');

1 row created.

SQL> insert into bowie_ro values (2, 'Ziggy');

1 row created.

SQL> insert into bowie_ro values (3, 'Major Tom');

1 row created.

SQL> commit;

Commit complete.

SQL> rename bowie_ro to bowie_temp;

Table renamed.

SQL> create materialized view bowie_ro
2 refresh on demand complete
3 as select * from bowie_temp;

Materialized view created.

SQL> drop table bowie_temp;

Table dropped.

SQL> select * from bowie_ro;

---------- --------------------
1 Bowie
2 Ziggy
3 Major Tom

SQL> insert into bowie_ro values (4, 'Thin White Duke');
insert into bowie_ro values (4, 'Thin White Duke')
ERROR at line 1:
ORA-01732: data manipulation operation not legal on this view

A few grants over here and a few possible FKs to change over there and the transformation is complete without the read-only tablespace.

Yes it's pretty simple and a bit of a cheat but the cat's kinda been skinned all the same ...


Richard Foote

Wed Jan 18, 06:07:00 AM EST  

Blogger David Aldridge said....

Ah, but partition-based DDL still works on materialized views -- you have to "alter table bowie_ro disable table lock" as well.

Wed Jan 18, 08:23:00 AM EST  

Blogger Gleisson Henrique said....

HJR has some valid points as some of the other readers too. However, I think we are forgetting one thing. Let's give OraQA a chance. On the about page of OraQA, you can see that all questions will be revised before posted, so maybe with a good moderator we can get good stuff (I believe the moderator is Eddie himself).

On the note of "having too many resources available", I will ask just one question. How many of us have our own blog /website with stuff about oracle and/or related technology? And why? Because we believe it's something that is not out there yet but it should be and maybe can be useful for somebody googling for answers. If our blog or website helps one person and one person only it was worth setting it up and maintaining it.

I have over 30 blogs/websites on my bookmark some I read everyday, some I read from time to time. Do I think it’s too many? No, I actually think it’s too few. There are hundreds of technologists who could be publishing material on the web sharing their knowledge.

Wed Jan 18, 09:19:00 AM EST  

Blogger Bob B said....


Nice strawman attack. Using the worst possible examples (e.g. rebuilding indexes = better performance) to "prove" your point and then furthering it up with a "throw out information" thing.

Look at what I wrote again. I was talking about people in the "I Suck Phase" of Oracle Development. Discussion and filler around an actual solution is just getting in the way of them getting out of the "I Suck Phase". Once they get out of that phase, the discussion and filler can be extremely valuable, but not until then.

Quite a leap from not learning from the information to throwing out the information. Someone can read and memorize information as much as they want, but they haven't truly learned it until they can pull it out and apply it when the situation arises. This takes practice, ie experience. Would you put someone in charge of your Oracle db backup and recovery who has read all about Oracle backup and recovery, but never touched the Oracle database? They've certainly learned the information, no?

PS - I logged in for my previous post, but blogger posted it as anonymous - weird ...

Wed Jan 18, 10:05:00 AM EST  

Blogger Joel Garry said....

How many of us have our own blog /website with stuff about oracle and/or related technology? And why?
The only reason I have a blog is to be able to post to other blogs. <But you knew that, right? >/sarcasm

I have several websites, each created to solve a particular problem. But that creates a problem - how does someone else get to the site? Google is a bad answer in this case, note Howard's issue about evaluating veracity. Aggregators? Still too immature of a technology to judge.

Look up dialectic in the dictionary. That is what asktom and usenet and Jonathan Lewis and some fora have - a dynamic interaction to determine the veracity of assertions. Usenet came up with FAQ's long ago. They are quite useful when limited to a relatively small data set. But with Oracle, you wind up with too much data, so see orafaq - I think it has both good and bad points, but compare it to the original Oracle faqs of years ago.

I'm not so radical as Howard seems to be on this - I think it is a Good Thing that Tom actively supports germination of community support. Good may well come out of the hacks required to make this work (somewhere on there is a hacks collection). One of the strong points of the web is it has generalized linking, so there can be a spectrum of ways to link, from very hierarchical (like blogging) to completely network. And, like someone famous said about usenet, it is a lot like an elephant with diarrhea.

What is a weed but a flower in the wrong place?

Wed Jan 18, 10:13:00 AM EST  

Blogger Howard J. Rogers said....

It wasn't a strawman. You stated that "I think its rare that anybody for any skill has ever learned much from reading", and in a nutshell that's the stupidest statement I've read in a long time. I merely tried to point out why.

Would I let someone who has never touched Oracle, but has read all about Oracle backup and recovery, recover my database? Assuming no-one else was available??

Of course I would! I stand a lot more chance of getting my data back that way then letting someone try it who has never bothered to do the reading because he feels that nobody "has ever learned much from reading".

Would I prefer someone who, having done the reading, took the time to experiment for themselves on play systems, to try things out, to do practical dry runs? Of course, too. But underpinning their practical experiments would be a sound understanding of the *facts*, not a two-line reductio ad absurdum of those facts.

I have no idea what the "I suck phase" is, let alone what being in it is supposed to be like. There is a body of technical information about which we all start 100% ignorant, and hopefully which we all end up knowing more than 50% of. I happen to believe going from 0 to 50 is a lot easier without having to unlearn later the crappy and misleading simplifications which you encounter early on.

If you are suggesting that newbies are too thick or too delicate to confront challenging concepts in a rounded fashion, then I believe you are underestimating them enormously. I have more respect for my students than that, and more faith in their abilities.

Wed Jan 18, 04:43:00 PM EST  

Blogger Teko said....

I think Richard Foote idea is interesting, but only if the dropped table is not in relation with some other table (foreign keys)

Buy the way, Q&A is interesting site. It is not a forum, it is not a blog, but for sure it is place where you can learn something either from the answers or from the posts.


Wed Jan 18, 05:35:00 PM EST  

Anonymous Scot said....

Two observations.

1. It just struck me while reading this thread: Howard is the Terrel Owens of the Oracle community.

2. In a sense, for the older questions, has evolved into a wiki / blog / forum hybrid. Starts off with an initial question, with only certain ones being accepted as topics to discuss to begin with. Many people post and discuss with one another and the moderator, and collectively the contributions form the final content, which gets clarified and later updated as Oracle versions change.

Wed Jan 18, 10:00:00 PM EST  

Anonymous Anonymous said....

Can someone please help the rest of us understand why DISABLE VALIDATE makes the table immune to DML?

Wed Jan 18, 10:13:00 PM EST  

Blogger Howard J. Rogers said....

Howard is the Terrel Owens of the Oracle community.

More ad hominem, I see. You mean, as per that I "have a history of bickering with teammates and coaches, and is known to put self promotion ahead of the best interest of the team."

It's a neat piece of circular argument, because I can't dispute the assertion without making myself into precisely the argumentative person you think I am!

All I will say is that I argue this particular point because I have spent nearly 20 years teaching people IT, and I am gobsmacked that anyone thinks it can be taught with two-line summaries that are wrong. I argue it because 2-line summaries distort and mislead. I argue it because newbies will make mistakes if they follow this approach. And I do it precisely because I want the Oracle community to prosper and improve, not mire itself into a lot of silly myths and errors.

What you think of me, I don't care. But to say that I put my own ego ahead of what is good for the Oracle community.. that I dispute with knobs on. It is absolutely the well-being of the newbie section of the Oracle community I have in mind in all of this.

Wed Jan 18, 11:11:00 PM EST  

Blogger Howard J. Rogers said....

Can someone please help the rest of us understand why DISABLE VALIDATE makes the table immune to DML?

Because if a constraint is disabled, it is not enforced... yet you want to "validate" it -in other words, you want to ensure that data conforms to the constraint's requirements.

So if you want to make sure the data conforms to the constraint's requirements, but you won't let me check each data change as it is made to see that it conforms... I have only one choice, which is to lock the table out from all DML attempts.

It was invented in 8i to allow people to drop indexes in order to re-create them. Dropping an index that is associated with a constraint implies disabling the constraint... but that would then leave the table open to "bad" DML. The new constraint state was invented to allow you to disable the constraint enough to let you drop the index associated with it, but not permit dodgy DML to take place.

Wed Jan 18, 11:55:00 PM EST  

Blogger Howard J. Rogers said....

My earlier reference should have been

I think the .aspx dropped off.

Thu Jan 19, 12:02:00 AM EST  

Anonymous Francois said....

I think that a reading team, approving the proposals before publishing them, would be a good idea.

Thu Jan 19, 05:51:00 AM EST  

Blogger Joel Garry said....

I think that a reading team, approving the proposals before publishing them, would be a good idea.
You've just described a moderated usenet newsgroup. I could point at one that demonstrates how bad that idea can be, but I don't want to support it by publicizing it here.

Or a cabal.

Thu Jan 19, 09:51:00 AM EST  

Blogger Bob B said....

"I Suck Phase" is when lack of knowledge prevents you from doing something. A sign of leaving this phase is when you're trying to choose between two or more options of doing something versus trying to figure out how to do it.

Our debate over learning is semantic. To me learning is about owning the information. Getting the information and putting it into your head is memorization. Having ample reading material is critical for memorization, but experience and practice are needed for learning.

I don't think "newbies" are dumb. What I do think is that we need to remove/destroy any barriers in the way of what they're trying to do.

Advice is only bad if it doesn't meet the requirements of the advisee. On Edie's Q&A you mention that the read only table answer is flat wrong. Then, here, you point out one case where it is 100% the correct solution: when recreating an index.

Lets say I have an existing table that needs a new constraint. I want to fix the current data and then put the table in read only mode while I apply the new constraint. I look around online and find a forum about it that gives the answer on Edie's page that is then refuted the next line. At this point, I'll either move on to a different search result, or, if I'm extremely patient, read on and on and get to the part where it says it fits my case directly. Or, I could hit upon Edie's site, find the answer, try it out and see that it works. Yes, the latter case could be problematic, but being able to do *something* is better than doing absolutely nothing.

WV: warftu

Thu Jan 19, 10:41:00 AM EST  

Blogger kevin lidh said....

The concept of two-line answers to complex questions reminds me of Oracle's "Two-Day DBA." Intentionally or not, it implies that it's possible to learn something complex without a great deal of time or effort (or discussion). Personally, I like following discussion threads because a great deal of time the truth is in the middle.

Thu Jan 19, 11:12:00 AM EST  

Blogger Howard J. Rogers said....


you point out one case where it is 100% the correct solution: when recreating an index.

So when all else fails, distort everything I actually wrote?

I must have missed the bit in the question published on Eddie's site where it asks "How can I drop an index when it is used to enforce a constraint without opening the table to the insertion of flawed data". I thought the question published was "How do I make a table read only".

I think you will find those two questions entirely different, and the answers to the one do not imply that it is "one case" of being an answer to the other.

Your elaborate example only goes to make the same point I started out making a long time ago: the answer to a question depends on why you are seeking the answer. If you are wondering about how to apply a constraint, then a temporary appearance of read-onliness might well be sufficient for your purposes. If you are trying to satisfy a tax inspector that account records have not been *capable* of being altered, then a temporary appearance of read-onliness will be insufficient for your needs.

Are you telling me that I missed the bit on Eddie's site where it discusses the difference?

Advice is only bad if it doesn't meet the requirements of the advisee

Thank you. My original point precisely. You'll have to help me out though: where in Eddie's Q&A style framework is there ANY attempt to discover the advisee's requirements, or to address matters in such a way that a range of requirements is covered?

Short answer is, it doesn't. It gives a 'one size fits all' answer to an un-elaborated question. It's that reductionism, of complex issues requiring an assessment of business needs, to trite one-liners that is the entire problem here.

Kevin: Thank you. Exactly what I've been trying to say.

Thu Jan 19, 04:43:00 PM EST  

Anonymous Tom said....

Scot's remark about Howard Rogers("Howard is the Terrel Owens of the Oracle community.")is uncalled for. Anybody who has spent any time at all reading Howard's writings knows that he is passionate about giving correct advice. People like me who have only scratched the surface of things Oracle know that when we read something Howard has written, we don't have to wonder whether or not it is so. I spend a lot of time on Howard's site (along with others whose advice I have come to trust) and I appreciate the time and effort he puts into it. Disagree with him if you must but don't resort to name-calling.

Mon Jan 23, 11:11:00 AM EST  

Anonymous Anonymous said....

Sorry but the mantra is don't trust prove it yourself. Howard is just as likely as many others who have been doing oracle for many years to be wrong. There is a difference however between many people who admit mistakes and other people that rarely if ever own up to that level of honesty.

Mon Jan 23, 04:49:00 PM EST  

Anonymous Scot said....

I think my Terrel Owens comment was misunderstood, and while I apologize for any hard feelings, I still think the analogy holds true.

It was not intended as an insult or a compliment; simply an observation that has helped me see Howard’s writing in a new light. I didn’t know about the link and reference that Howard quoted. However, if you continue reading on that page you find other people who are fans of Terrel posting in support of him.

He has a somewhat raw style that doesn't sugar coat anything, and he doesn’t shy away from speaking his mind. He is strongly opinionated, and has an ego that is born out of self confidence. Naturally all of these things rub many people the wrong way, especially if his comments are taken personally.

Above all else, he is a polarizing figure who is one of the best in the world at his profession.

Wed Jan 25, 10:55:00 AM EST  

Anonymous wokiko said....

This looks like a great reference site. It's sort of like the youtube how-to videos. However, my favorate q&a site is because you actually get paid to answer questions. Like previously stated there are thousands of q&a sites but the problem with most is that people like to respond to almost anything regardless if they know the answer or not. There are no incentives for correct answers.

Wed Jan 09, 11:53:00 PM EST  

Blogger laptop battery said....

What you said is I thought of .Here is some of new battery fpr your laptop.

Sat Aug 09, 07:30:00 AM EDT  


<< Home