A Base issue cover-up?

[DACM]

I consider it a distinct privilege to participate in this and other community forums, predominately in support of Base. The moderators have endured my sometimes outspoken critique of Base in the process of urging Base users to adopt a reliable database configuration for production data. I simply joined the campaign of r4zoli, Sliderule, Villeroy and many others as we've sought to maintain a proper balance between transparency and Base/Office promotion -- as avid users ourselves. As a result, any 'seasoned' Base user is now keenly-aware of the Base instabilities associated with storing database-files internally, within the Base .odb file, otherwise known as 'embedded database' files. We continue to encounter newer/un-aware users that have fallen victim, and find themselves desperately begging for help with critical data recovery. For instance, here's today's entry. Our support includes data-recovery, automated backups, configuration tutorials, data-migration tutorials, automated templates, examples, personalized user-help, etc. To that end, I've painstakingly generated detailed tutorials outlining each of these solutions (up through concurrent database access among multiple users in networked environments) to the official Apache OpenOffice.org Wiki.

I understand the nature of a Wiki (and the ability to 'undo' changes), but apparently those tutorials were removed recently by Eschmenk citing "Major rewrite of some confusing sections." Now I can appreciate any effort to clarify while condensing my prose, but this effort substantially eliminated the tutorials and all links to this forum (leaving all others intact), while further sanitizing the content of any reference to data-corruption or known Base instabilities/workarounds/wizards/etc. In other words, Eschmenk intentionally gutted the content and all references to easy/automated solutions, while effectively eliminating all motivation to avoid 'embedded database' files. What's left is a confusing and suspicious fraction of the original tutorials and information.

This is troubling because we can no longer presume that new and inexperienced Base users have encountered data-corruption due to a lack of exposure to information channels. If Eschmenk represents an insider, then we we will soon be encountering victims who have lost data due to a lack of motivation/awareness, due to the sanitized nature of the propaganda allowed through official channels.

Thankfully, we enjoy much greater transparency in this forum which serves to overcome the intentional misrepresentation of the facts by insiders. But it still hurts to realize that we have such [apparent and blatant] dishonesty in the form of a cover-up extant in the open-source community.

See related excerpt (Sep 03, 2013) from: FAQ (Base) - Apache OpenOffice Wiki.pdf

Edit: I was finally able to access the Wiki History. I'll save a link here for achiving puproses: Differences link or this PDF . ...

[crusader]

DACM, as an occasional Base user, I am curious: are the problems you have identified related to Apache OpenOffice only or is LibreOffice's Base also affected?

[Villeroy]

DACM, as an occasional Base user, I am curious: are the problems you have identified related to Apache OpenOffice only or is LibreOffice's Base also affected?

LO Base is (almost) identical to AOO Base.
http://nabble.documentfoundation.org/Ba ... 55060.html

[Villeroy]

I feel personally addressed by the following paragraph:

When are changes to a database recorded? Are any temporary files used?

As implemented by Base, changes made to a HSQL database are silently recorded directly to the database files. Even when the database is embedded in an .odb file, any changes are written directly to the .odb file. It is not true that changes are written to temporary files (or held in memory) until the user exits from base, and only then written to the .odb file, despite what some people have guessed.

Well, it was not guessing which led me to my own conclusion.
Open some editable record of an embedded database.
Notice the temporary files in a temporary sub-directory and notice the database document itself.
Edit/insert/delete some records.
The temp file changes, the document doesn't.
Close your table, query or form.
Now the document is updated.
So my own misconception was that the document's db updates on closing. In fact it updates when you close the editing tool (grid view or form document).
Practical relevance: None. Losing embedded databases is still a high rated problem. My own forms are loaded on 3 work places from 8:00 am until 18:00 pm. They are closed by accident or by system shutdown. Both, the forms and the database, are not embedded and work flawlessly since 30 months.

[Hagar Delest]

For the record, RoryOF raised the issue on the dev mailing list, here is the link: http://www.mail-archive.com/dev@openoff ... 12122.html
Let's see how it develops. I'm not very pleased with the beginning...

[lgusaas]

I understand the nature of a Wiki (and the ability to 'undo' changes), but apparently those tutorials were removed recently by Eschmenk citing "Major rewrite of some confusing sections." Now I can appreciate any effort to clarify while condensing my prose, but this effort substantially eliminated the tutorials and all links to this forum (leaving all others intact), while further sanitizing the content of any reference to data-corruption or known Base instabilities/workarounds/wizards/etc. In other words, Eschmenk intentionally gutted the content and all references to easy/automated solutions, while effectively eliminating all motivation to avoid 'embedded database' files. What's left is a confusing and suspicious fraction of the original tutorials and information. .

Undo the changes. If they are changed again, start discussion. Eschmenk seems to be new to the wiki. I don't see any other contributions

[pescetti]

Undo the changes. If they are changed again, start discussion. Eschmenk seems to be new to the wiki. I don't see any other contributions

Exactly. I cannot look at the page history now (intermittent network problems with the wiki today). However, I tend to believe that the work of several long-time contributors should not be deleted without discussion. Post a notice to "Eschmenk" (whoever he/she is; for sure he is not a regular contributor) in the talk page, and make sure that the wiki page is on topic and accurately reflects reality.

If it describes bugs, fine. We are here to solve bugs, not to hide them. I can't believe that someone really thinks that the project would have an interest in hiding Base bugs. It is true that bugs in Base historically have received less attention, for various reasons (don't forget that Base is also substantially newer than other modules); but words like "cover-up" really do not fit with the current totally open governance we have at Apache.

[DACM]

Okay. I think based on the replies we've received here and through the Apache dev mailing list, I'm happy with treating this as an isolated incident. And I'm sorry for suggesting otherwise. Please accept my sincerest apologies throughout the Apache governance. I'm now confident that this can be resolved through a simple discussion with Eschmenk.

Please understand that this allegation came on the heals of several such incidents in the LIbreOffice camp (such as Villeroy linked above), to include a leadership-driven campaign against Java that's gone unchallenged and now threatens to create another embedded-database debacle based on Firebird.

Otherwise, the user-community has stumbled upon a good alternative to 'embedded databases' which might be of interest to the Apache devs. It's backwards compatible with legacy embedded database support; it's cross-platform; easily-portable; and it's deemed sufficient by the user community. Simply download a macro-enhanced template for a preview of Villeroy's efforts, combined with my own ideas on the subject. You'll be using HSQLDB 2.3.0 (internally licensed) with Base in a matter of seconds. The devs could perfect the idea while performing some clean-up work associated with DDL-tasks for JDBC/ODBC connected engines. And we've outlined an acceptable single-file database solution for distribution and self-extraction purposes.
...

[DiGro]

I've put up a notice voor user Eschmenk to contact DACM on the involved wikipage and the Talk:page.

If not responded in 7 days I will remove the notice from the wikipage (If noone beats me to it

It is really sad to loose this valuable content

[Greengiant224]

What on earth is going on here?

DACM and all of the users (including me ! !) who have contributed to the setting up and running of single file/multi user
database connections must be mortified at the apparent 'deletions' of the 'work'
that has been put in at the wiki page.

Perhaps this may be an opportune moment to put all this data into a.pdf file linked from this forum.

Maybe, just maybe Eschmenk may have a change of heart after realising he/she has made a big mistake.
I won't be holding my breath.

Greengiant224

[RoryOF]

The view on the developer mailing list was that Eschmenk should be asked to explain and a suitable solution evolved if possible. It was my feeling that they preferred that this course should be followed first, leaving the door open for other choices if that did not work.

But a .pdf/ tutorials can certainly be hosted on this Forum with whatever security may be deemed necessary,

[DACM]

If not responded in 7 days...

Undo the changes.

I think it will come down to undo-ing the changes, particularly if Eschmenk doesn't respond in a reasonable time-frame. A rollback should be a matter of a few 'undo' clicks.

I certainly don't want to give anyone the impression that I'm opposed to edits, clarifications, consolidations, etc. to my Wiki content. I take no ownership beyond a concern for the necessary information. My concern in this case was for the alleged agenda behind these particular edits, combined with the fact that they were undoubtedly crafted by someone with significant experience with Base (despite some inaccuracies).

So absent discussion with Eschmenk, I'll plan to rollback the changes in a week -- while taking the brevity and clarifications contributed by Eschmenk under advisement.

In light of this incident and the nature of a Wiki, I do plan to generate the tutorials in this forum as the primary reference.

[Hagar Delest]

I certainly don't want to give anyone the impression that I'm opposed to edits, clarifications, consolidations, etc. to my Wiki content. I take no ownership beyond a concern for the necessary information. My concern in this case was for the alleged agenda behind these particular edits, combined with the fact that they were undoubtedly crafted by someone with significant experience with Base (despite some inaccuracies).

+1.

I know nothing about Base. But looking at the edits, it was obvious that there was some agenda behind. This was not done by a newbie as said on the dev mailing list. The incident has been rather downplayed on th edev list IMHO, which is clearly a sign of uneasiness, in addition with the very few replies on that discussion.

[DACM]

Undo the changes.

A rollback should be a matter of a few 'undo' clicks.

Done.


...

[pescetti]

looking at the edits, it was obvious that there was some agenda behind. This was not done by a newbie as said on the dev mailing list. The incident has been rather downplayed on th edev list IMHO, which is clearly a sign of uneasiness, in addition with the very few replies on that discussion.

If the guy had an agenda, it was his personal agenda, or his company/whatever agenda, but definitely not the OpenOffice agenda (which does not exist, or better, is formed daily through the messages on the public dev list). I'm happy that a simple revert solved it, but the scarce reaction was simply due to the fact that we have no proof of organized attempts to change the OpenOffice wiki (and by he way the wiki was unavailable during the discussion, so this didn't help understanding the situation).

If you see anything that looks like an organized attempt to modify a lot of content in some strange way in the OpenOffice wiki, please report it to the dev list as you did. For sure we don't want this to happen without a preliminary open discussion on the dev list.

[Hagar Delest]

I'm pretty sure there is no agenda from the AOO project [to make such changes I mean].
BUT what I found rather strange is the lack of reaction, that's all.