Original Article

My comments...

"Microsoft bug corrupts records"

A flaw in Microsoft's Access database,⁽¹⁾ which causes changes on one Access database record to be saved on another and can leave users unaware that their records have been corrupted, is traceable to Microsoft's Jet database engine.⁽²⁾

The bug strikes when, in a long form⁽³⁾
with dozens of records, a user deletes a record set⁽⁴⁾
that appears early on in the list, then makes changes to a subsequent record and saves that change. The flaw occurs when users view Access using its Form View, then use the Combo Box feature to edit the record.⁽⁵⁾

⁽³⁾ -The form length is irrelevant (the number of records is not)

⁽⁴⁾ -"deletes a record set"??? What does that mean? No DAO sets are deleted.

⁽⁵⁾ -The bug has nothing to do with combo boxes, and certainly doesn't require comboboxes to be involved in the edit

Essentially, under those conditions, Access 97, Access 95 and Access 2.0 do not recognize that the first record was deleted, so the rest of the data is corrupted.⁽⁶⁾

⁽⁶⁾ -What do they mean by 'the rest of the data'? The remaining data in the db? the table? Only the edited record is corrupted. What this means to the db at large depends on its use.

"The fact that Access displays one record but writes changes to another undermines its credibility," said Allen Browne, an Australia-based Access developer who posted information on his web site.⁽⁷⁾

⁽⁷⁾ -And had the authors taken the time to read their article back to Allen when they were done, this mess could have been avoided entirely.

The bug is related to Access's interaction with its underlying database engine, the Jet engine, occurring because of a pointer mishandling between the Jet engine and the Combo Box feature.⁽⁸⁾

⁽⁸⁾ -Again, it has nothing to do with combo boxes, which aren't a feature anyway. They are standard controls.

Because it is specific to the Combo Box, an Access feature,⁽⁹⁾
the array of other Microsoft technologies that use Jet - such as Visual Basic and Visual C++ - are not affected according to a study by the InfoWorld Test Center.⁽¹⁰⁾

⁽⁹⁾ -no its not [it is] a control

⁽¹⁰⁾ -if their test center can't even determine the relevance of comboboxes in the primary technology, one should not trust their opinion on other technologies

Microsoft is preparing a fix that will be included in Service Release 2 for Office, which will provide corrected pointer handling between the Combo Box and the Jet engine in Access. This fix is expected to be viable for all of the Access versions affected.⁽¹¹⁾ Microsoft could not provide a timetable for the release's availability.

⁽¹¹⁾ -Microsoft will not release a patch for 2.0, or 95, if past experience is any guide. They will just provide a a 97 patch (that can only be used with 2.0 and 95 databases when they are opened under 97 itself.

Microsoft issued a workaround for the bug last week. Microsoft suggests users go to Access's Design View, right click on the Combo Box, type me.requery into the dialog box, and save it.⁽¹²⁾

⁽¹²⁾ -Microsoft clearly suggested nothing of the sort, especially since there is no dialog box that pops up when the user right-clicks a combo-box in design mode. There is just a context menu. Surely the authors could have at least tried their own suggestion. Even if they felt they were incompetent to judge whether the method is indeed a fix, they could at least have seen whether it was possible to perform (it isn't).

Database rival Oracle wasted no time in playing up the Microsoft bug, posting on its own web site tools to speed migration from Access to Oracle8 and Oracle Lite.⁽¹³⁾

⁽¹³⁾ -These last two paragraphs, the first about a company (Oracle) not involved with the bug directly and the second covering just contact details, are the only ones without significant errors!

Microsoft Corp., in Redmond, Wash., can be reached at http://www.microsoft.com. Allen Browne's web site can be found at http://allenbrowne.com/homepage.html.⁽¹³⁾