The unnecessary censorship discussion post...

Update: Before posting this, I had made an effort to verify that this was not a technical problem. However, Lenz from the MySQL Community team has now replied that indeed Zak Greant's RSS feed had changed it's URL address (see comments), which is why new posts did not appear. So rather than telling anything about Planet MySQL, this post is good evidence of my and others mistrust towards it - thankfully still unfounded. (The roots of this is also discussed in the comments.)

NOTE: Heading of this post was subsequently changed.


Not that I would be surprised of anything anymore, but still.

Ok, so ever since I started working for Monty, we have become an outlet for people that have complaints, suspicions, anything against the official Sun/MySQL organization that people may have. So it is not that we like to complain, things just come to us. (And most of it stays out of the public.)

During the summer there were 2 incidents that people suspected Planet MySQL being censored. The first one turned out to be a possible problem in browser cache. (Or maybe something was censored and then put back?) For the second one: "Wow, I'm not sure it is a coincidence. An internal Sun/MySQL employee posted something in support of Open Database Alliance on sun.com right after its announcement, which I read, and within 2 hours it had disappeared." ...it is likely that planet.mysql.com was not censored, but rather the employee was told to remove his blog post completely, I know that happened every now and then. (And it is a different kind of censorship, say what you want but at least it is commonly accepted that employees should just shut up when their boss tells them to.)

But now I notice that Zak Greant's latest post where he disagrees with Mårten's open letter to the EU is not on Planet MySQL. It is properly tagged with MySQL, like all other his posts that are on planet.mysql.com.

So disagreeing with Mårten is not allowed now? Or what is the problem? (Mårten's letter got quite many + votes, so apparently the topic itself is popular and newsworthy, just the opinion is different.)

If this is censorship I'm seeing, I'm a bit surprised. A Planet is only passively syndicating blogs from other sources. It is obvious to anyone this is not something Sun or MySQL is writing. As long as this is the case, then everyone understands the original author is responsible for what he writes, not the Planet site. But if you start censoring the Planet itself, suddenly there is an active publisher role. Then the next time there is a post someone doesn't like, you may have to censor that too, because now you can't claim that the Planet is just passively syndicating other blogs without no editorial control from its maintainer - you suddenly have editorial responsibility.

Do other Planets ever censor anything? Ever?

Note that I would also understand if someone was expelled from a planet feed, such as for spamming off topic posts. But being selective on a post-by-post basis is purely censorship - or put more mildly "active editorial engagement", which is not what a Planet website should do.

Bryan (not verified)

Fri, 2009-10-16 18:50

I guess the ODBA needs to get a planet up soon.

Have considered that this could be a technical problem, not malice? I just checked the Planet MySQL entry for Zak's blog, it was pointing to http://zak.greant.com/index.rdf as the RSS feed, which seems to no longer exist (last post received from there was May 20th). It has moved to http://zak.greant.com/feed/ it seems, but he did not inform us about this change. I've now fixed the entry, so his recent postings tagged with "mysql" should appear shortly.

Next time I'd appreciate if you would get in touch with us directly instead of raising such conspiracy theories in public. Thanks!

Hi Lenz

Yes, I did have the peace of mind to actually check with Zak, and he probably didn't remember this change, since he couldn't think of a reason why he hadn't appeared on Planet MySQL. Since his last MySQL post was on planet.mysql.com, there was no reason to suspect technical failure either.

I have to admit, due to the sensorship that happened at the user conference last year, it unfortunately reflects badly on Planet MySQL too. Like with the above 2 examples, the thought "I'm sure Sun would never do that" doesn't really come to people's minds anymore, and I'm guilty of lacking that trust too. But I'm glad to hear Planet MySQL still has a clean record!

Thanks for clarifying, I'll update the post with your info.

"due to the sensorship that happened at the user conference last year"

What censorship are you referring to in this case? I'm curious because there is a lesson here to any company that wants to build a relationship with the community.

Despite the fact that this incident was a red herring, your reaction shows that ill will is easy to build and it's very hard to build back a reputation when it is poisoned. This is especially true when there is a significant power difference in that relationship, as in Sun's participation in the community.

Hopefully the new stewards of MySQL and also other companies such as your own can learn from past mistakes and build a more trusting community.

Hi Ryan

I don't know if "censorship" is the right word, but I don't have a better one. At the MySQL conference 2009, some speakers that have historically been popular, were not selected to speak at all. We were assured that the selection was fair, but certainly this was suspicious, since it is unlikely that for instance Percona, who had submitted several proposals, including by the authors of arguably the best MySQL book ever written, were not selected for even one talk.

This strategy also backfired badly for Sun, because Percona then announced their own conference to happen in parallel with the MySQL conference. Eventually Sun was forced to fix the situation by allowing Percona to speak at the main conference, as well as hosting the competing Percona conference and paying for the premises :-)

I've since gotten confirmation that the exclusion of certain speakers who happen to be MySQL competitors was (of course) not a coincidence. (Others than Percona never made it to the conference.)

This post from Baron Schwartz provides some historical details and further links.

I'm the first to admit that this incident did influence me negatively when it comes to trusting MySQL management (admittedly some of them have since left) and influences my judgement nowadays. So also when posting this post, I did of course check some facts before raising an alarm, but actually contacting Sun employees didn't occur to me as necessary. And of course, when you've been lied to once...

But I'm still very happy to have it confirmed (and absolutely want to believe) that planet.mysql.com is not censored and we still don't have a single incident where it would have not been a false alarm.

If there is anything positive to this, I hope talking about this may at least clear the air a little bit. Like I explain in my blog post, I've gotten 2 mails during the summer where others have suspected censorship, but that were not. (So we never raised alarm of them.)

Hi Henrik,

thanks for posting the clarification. Unfortunately the subject line on Planet MySQL won't update to include the "(Update:NO)" part if you exclude this post in the RSS feed going to it :)

One thing that is on our TODO for Planet is to allow existing users to maintain their feeds by themselves. This will make it easier for individuals to check and update their own info.

Until then, the best way to check the status is to get in touch with us, e.g. sending an email or using the contact form that is linked from the Planet front page.

Have a nice weekend!

Bob tables (not verified)

Fri, 2009-10-16 20:35

Mysql planet doesn't censor, but I do. I run the mysql planet feed through yahoo pipes to exclude the flamebait and beginner level like stuff. I'm afraid I'm going to have to add your feed to the banned list. This is just too reckless of an accusation.

There is a "Feedback" link on http://planet.mysql.com/ which you should use.

Note that tagging isn't enough; a post also has to say the word "mysql" in it somewhere.

Also note that some people may think there is censorship because there is a very log polling interval -- I think it's about an hour (and they reduce it for the MySQL Conference).

As for the conference "censorship", there was in fact a delay processing Percona's entries, and Percona took the delay to mean rejection. (it didn't help that there was an actual accidental rejection letter sent to other people, including me, but an explanation was sent out too....but the Percona folks didn't accept that explanation).

Percona didn't even bother to submit anything to MySQL Camp, which is a free conference that *I* ran, and *I* chose the submissions for, and I can tell you I did not censor anything. I was *nice* enough to *save* 2 slots for Percona (because I didn't want to be accused of censorship) and it was like *pulling teeth* to get them to actually say what they were going to talk about -- even though I saved them spaces!!

So instead of waiting with their delayed proposals, submitting to MySQLCamp, or perhaps *asking* MySQL why their proposals were delayed, they went ahead and made their own competing free conference. I asked them why they didn't submit to MySQL Camp and the reply I got was "we want it to be branded with Percona".

They also made a big deal about how Percona conference was free, and MySQL wasn't, and completely ignored MySQL Camp, which had been announced even before the call for proposals for the MySQL Conference.

Also check out this thread:

http://groups.google.com/group/opensqlcamp/browse_thread/thread/a2f1188…

Percona asks if OpenSQLCamp 2009 will accept Percona presentations....

....it was pointed out that there was a lack of Percona proposals for OpenSQLCamp 2008

....then Percona asks "what should we talk about?"

....somehow they were wondering what kinds of proposals were appropriate for the conference. Which is odd....why can't be bothered to look at a web page, like everyone else?

And as much as we all have had issues with MySQL as a company, it really looks to me like Percona is going the same way:

http://www.percona.com/docs/wiki/release:start#comment__7638c96f08aec1b…

that link is a reply to someone asking about a feature, where Percona says "we're concentrating on what paying customers want, not what the community wants".

Whenever anyone says anything critical about Percona or their software, they immediately jump into defensive mode, as opposed to accepting the criticism as valid. On this page, I'm accused of FUD:

http://www.pythian.com/news/3241/concerns-and-what-does-not-work-in-xtr…

the post (me) says: I don't know why the ibdata1 file is changed, but that's kinda scary.

vadim says (comment 1): it doesn't change ibdata1, it works the same as innodb hot backup (ibbackup).

I reply (comment 5): it doesn't work the same, at the very least because xtrabackup requires the MySQL user.

baron replies (comment 7): read the source code, or prove that it requires the MySQL user.

[How rude! If MySQL *ever* gave that reply, there would be blog posts all over the place!]

I reply (comment 8): we have done what you say. here's proof.

Baron replies (comment 9): he misunderstood something, he apologizes, he'll ask Vadim.

[even reading it over now, I can't imagine how Baron was confused...but even if he was, Vadim insisted in comment 1 that it wasn't different, when I specifically said it was in the post!]

Vadim replies (comment 10): Oh, it does write to the ibdata1 file. They have now fixed that issue in the next release.

----------
There's more, and Andrew lambasts *me* because I assume that "write access is required" means that the data file is written to. I don't think my conclusion is unreasonable, and as soon as it was explained to me I updated the post, but that whole thread is ridiculous and could have been cleared up very quickly if Vadim had been honest in comment #1 -- he said that I was wrong and xtrabackup works exactly as ibbackup, and that is not the case, as he stated in comment #10.

-----------------------------

I like Percona's software, but I worry that they are being hypocritical when they criticize MySQL, because they are doing some of the same things (prioritizing features for paid customers) they didn't like about MySQL, and they seem to be more evil (like setting up Percona Conference even though MySQL Camp already existed). :(

Hi Sheeri

About how planet.mysql.com works: As you can see in my other comments, I know that MySQL has to appear in the article, I know there is a delay, and I had checked all I could think needed to be checked. Unfortunately I got incorrect information back, so now I look like an idiot. That's ok, I personally like discussing things in public, sometimes it means I'm the one who is wrong in public, I still think it is healthy for an Open Source community.

About what happened with the last MySQL conference: You are one of the nicest persons in our community, and therefore you probably also assume that others are nice people. I'm like that - I mean I assume that about others too. Just like you, I always believed it when I was told the speaker selection process was fair and there was no politics involved. I even tried to defend Sun on various blogs, since I believed that was the truth.

Imagine then how shocked I was to read the email from Mårten himself where he admitted that excluding certain speakers (Percona wasn't the only one) had been intentional, and trying to justify the reasons for it. (Besides that email, it has been confirmed to me outside of Sun too, not that the email was in any way unclear.)

I still suffer from that shock - as you have witnessed. We all knew MySQL management was always pulling into a direction with less Open Source, just to give one example where my values differ. But being lied to really hit me hard. I don't know what else to say.

As for the other points you have against Percona, I guess I wouldn't have attended MySQL Camp either after such treatment. The Camp is a good addon to the main conference, but if you've been specifically expelled from the main event... I would also be too proud to move my talks there. Our company didn't manage to send anyone to the latest Open SQL Camp either, so I can't really judge anyone else for not doing that, you of course can if you want.

About the other points, you certainly have a better style than either Percona or myself, and I hope to learn more from you how to behave in public. (And there is no sarcasm here.)

Eventually MySQL did reject the talks, but at the time that the "outrage" that Percona wasn't going to speak at the User's Conference, it was still a delay. The Percona conference was organized before the talks were actually rejected, which means the actual rejection were probably the reason they were rejected.

I know of 2 other speakers who were effectively "banned" from the conference; one because he had a few incidents of plagiarizing in his writings and presentations, and the other was banned because he lied to government officials to block a MySQL training from happening in his country, such that MySQL's visa application was denied.

I'd ban both those folks, and any legal team would advise that if you're thinking of legal action against a company, accepting talks to your conference sends a message of acceptance to the company, and can be used against you in a lawsuit.

I got the same information you did, and part of the "reason" for the delay was that there was a possible breach of contract on Percona's part -- but of course MySQL didn't want to say that because they were working on learning the facts first. Either there wasn't a breach or there wasn't enough evidence, because as far as I know there hasn't been a lawsuit filed.

(but this is also part of Percona's complex that MySQL was "bullying" them -- trying to get Percnoa to comply with the terms set in a signed contract is not bullying!)

(note that this is all stuff I've heard, and thus is legally considered hearsay)

Sheeri, I "lambasted" you because you made an over exaggerated claim - you claimed there was a direct write to the InnoDB data dictionary. If it had really been, as in your account here, that you were merely saying "the ibdata1 file is changing and I don't know why and this worries me" I would have found no reason to say anything - and really that exchange shows how difficult it is to have a serious technical conversation with you. Still, I and my colleagues are baffled how you can jump from "requires write access" to "oh noez direct write to the innodb data dictionary!!". When an experienced DBA is claiming such specific behavior there really should be some evidence.

The fact is you seem to *still* be claiming there is a write. Which is absolutely wrong, wrong, wrong. You should be embarrassed.

I am neither a Percona employee nor do I have no financial interest in Percona, however I do find posts like yours to unnecessarily sow mistrust in the community and add very little content.

Andrew -- I'm not claiming there's a write. I know there wasn't one.

What I'm saying is that it's *reasonable* for anyone to have assumed that there's a write, if *write access* is required.

As soon as I understood that there were no writes, it was just write access that was required, I updated the blog post and thanked people for the information.

However, I still believe that if something requires write access, it is *reasonable* to think that it could be writing.

(to be clear, I know *now* that there wasn't a write. I didn't know when I made the post.)

Yes, I made an assumption, and it was wrong, and I clearly stated that it was wrong when I found out it was wrong. But I also clearly explained that it looked like a write was being done because write access was needed, and the first response from Percona was "it works just like innodb hot backup" -- which I had said in the post that it did *not* work just like it.

If the first reply had been "oh, it requires write access but there wasn't a write," that would have been fine.

Henrik,
If you are implying that we (the community team) are in the business of removing blog posts that our bosses may not like, please note the following:
1) We don't do that;
2) Even if we wanted to do that, the way Planet is organized is "all or nothing". Either we remove a feed entirely, or we allow it. We can filter contents, but we do it to make sure that only MySQL related posts go to the planet. If I remove an entry, it will be reinserted 20 minutes later, when the feed reader refreshes the Planet. You can ask Arjen, who originally wrote the code. It still works that way. Not the best design if you want to censor something. For this reason, your post is still on Planet, although you have removed it from your feed. Removing it would require a deletion from the database.
3) In the case that you have mentioned about employees removing contents from their blogs, I can guarantee that we haven't been involved in such action, as you know very well, since you removed a post that someone asked you to, and Planet cached the original version online for a while after you changed the text of your blog. Eventually the Planet feed reader caught up with your blog and updated the public contents, but the censorship came from your side, not ours.
4) We have sometimes removed feeds from Planet for technical reasons, the authors were informed when that happened, and the feeds were restored when the technical problems were solved.

In case you have forgotten, there are several ways to contact us. We have email, IRC, Skype, phones, Twitter, Facebook, and some more. Talking with us is easy. Asking before shooting could have saved you the time of writing this unnecessary post.

Cheers

Giuseppe

... but my impression as a lurker on the MySQL internal channel of the community team on the MySQL internal IRC server is that we ocassionally had some technical hick-up during the past weeks. For example, I saw MySQL support team members asking why their postings did not show up.

If that happens to any of your postings, just ping the MySQL community team.

About the bookAbout this siteAcademicAccordAmazonAppleBeginnersBooksBuildBotBusiness modelsbzrCassandraCloudcloud computingclsCommunitycommunityleadershipsummitConsistencycoodiaryCopyrightCreative CommonscssDatabasesdataminingDatastaxDevOpsDistributed ConsensusDrizzleDrupalEconomyelectronEthicsEurovisionFacebookFrosconFunnyGaleraGISgithubGnomeGovernanceHandlerSocketHigh AvailabilityimpressionistimpressjsInkscapeInternetJavaScriptjsonKDEKubuntuLicensingLinuxMaidanMaker cultureMariaDBmarkdownMEAN stackMepSQLMicrosoftMobileMongoDBMontyProgramMusicMySQLMySQL ClusterNerdsNodeNoSQLNyrkiöodbaOpen ContentOpen SourceOpenSQLCampOracleOSConPAMPParkinsonPatentsPerconaperformancePersonalPhilosophyPHPPiratesPlanetDrupalPoliticsPostgreSQLPresalespresentationsPress releasesProgrammingRed HatReplicationSeveralninesSillySkySQLSolonStartupsSunSybaseSymbiansysbenchtalksTechnicalTechnologyThe making ofTransactionsTungstenTwitterUbuntuvolcanoWeb2.0WikipediaWork from HomexmlYouTube