Boy, David, you sure are digging.

>Jim,
>
>>>You reported a bug that reindex no longer removed CDX bloat. I proved your bug report to be incorrect, reindex always has removed CDX bloat.
>>
>>You did nothing of the sort!
>
>Rows 7 & 8 in the table of the Wiki article, directly refute your bug report. http://www.levelextreme.com/wconnect/wc.dll?FournierTransformation~2,15,652071

Maybe you can enlighten me as to how it is that rows 7/8 refuted my contention. Please bear in mind that my assertion was strictly and wholly about bloat caused by record insertion activity and you said in the article that such test results were to follow.

>
>>Your promised second report (quoting "Since the code below does not do any record insertions this analysis does not cover the secondary type of bloat. That will be covered by another set of test code posted in the near future." never did materialize
>
>The event in my family life that causes the link to the grid.org cancer research project to be included in my signature block prevented me from returning to the topic of CDX internals. By all means, if you are still interested in the topic go ahead and do the research and post it.

That truly is a sad thing. But I had done/shown ('published' is a tad pretentious) my research at the atart of the thread and feel no compelling need to redo it to obtain the same result.
You have had 2 years to finish it yourself. Maybe some day you should.
>
>>My my my, how a story changes over time!!!
>
>It does Jim, you seem to be now reversing the agreement you had in this message http://www.levelextreme.com/wconnect/wc.dll?FournierTransformation~2,15,653417

Not at all, David. Did you actually READ the message you cite above? I pronounced your work to be excellent and then noted that it was the second set of test results that were germane to my complaint and so awaited those results anxiously.
>
>>You were quite pleased with the rationalization for the doubling of the CDX size put forward by ChristofL at the time and happily accepted that it was all by design and to fix a/some bugs. It was all hypothesis and guessing and turned out top be a full crock of crap yet you persist with this double-speak!
>
>I acknowledged the doubling of size to be a bug. I also stated in the threads here that there are some positive benefits to changing the way B+Tree nodes split. Ulitmately though VFP is much more adversely affected by a doubling of size of the CDX for large tables and it was changed. I wish that a new setting had been added to the language so that we could control this on a table by table basis to help us optimize our databases.

Geez, did I miss that (acknowledgement as a bug)? Guess I did. Strange. And you continue to say here that it wasn't really a "bug" so much as it caused problems that were too onerous and so the VFP Team reversed it. Which is it man????

>
>Frankly I trust Christof's investigative research into the issue much more than I believe the "explanation" you offer in the Wiki article. Quoting you: I'll offer mine... when making changes that can affect TAG manipulation/accuracy the VFP Team facilitates its debugging by removing the 'compression' normally done so that they can better see the full story when needed. They simply forgot to turn on the compression again in VFP7SP1 once their TAG testing was finished. Sounds as plausible as any to me.

Believe me, I trust ChristofL far more than I trust myself on this sort of thing. I was simply offering another plausible reason for the bug. You see, if an analysis starts out on the premise that some bug fixes have resulted in changes to the construction of the CDX, and then you do find changes inside the CDX, it is a logical step to conclude that the changes now observable in the CDX are wholly attributable to the fixing of some bugs.
>
>>No, I feel that the incident effectively stopped further 'persecution' of myself.
>
>Or perhaps it just means that I've not seen you make any invalid bug reports of late. I'm an equal opportunity debunker.
>
>>That FabioL has contributed SEVERAL legitimate bug reports also makes no difference. You intend to shut him up and you will contine until that happens.
>
>Please point me to any criticisms I've made to any of his valid bug reports.
>
>>The operation was exposed with my indicent and it is now following exactly the same script as regards FabioL. But I suspect that his skin is thick enough to resist it too.
>>
>>Summarizing: I feel I survived the persecution and no longer feel any. You (the gang) have moved on to FabioL as I see it.
>
>You are just too paranoid. If I see something posted, by anyone, that is seriously flawed I will reply as I see appropriate.

Right, but you see no reason to apologize for inane behaviour on your part once the "seriously flawed" argument is found to be correct. Not a bad deal. Not my way of operating though.

Jim