Posted on 5 Comments

OSCON MySQL foo

Kaj writes:

We‘re happy with the quality [of 5.1]

  • More stable than 5.0 was four months after GA
  • A better MySQL 5.0 (thousands of small fixes)

This would be because 5.0 was a complete disaster in terms of GA quality. Peruse bugs.mysql.com, it’s all there. The number of open bugs around GA time, repeated API breakages (involving leaked symbols, inability to link against SSL, Postfix, and other apps/libraries), the number of (perhaps necessary but incompatible) behaviour changes in later updates, the number of regressions introduced by these may changes.

So, saying the current 5.1 quality is better than the 5.0 shambles is a) bogus and b) pure spin. Unworthy of a community rep. My opinion? Because all these things are public, a MySQL version can be considered stable when people “out there” say it is, not when any MySQL person says so, or an “official” statement is made. Really. Any spin statements merely hurts the community karma of those who write them. Kaj has a lot of good karma there, but he’s also been burning it off at a really high pace.

I like 5.1, it has quite a few interesting features, both small and large. However, I am still tracking a number of issues in 5.0 that need to be resolved. While the problems were filed for 5.0, they mostly exist in 5.1 as well. They are 5.0 bugs, but they do affect the 5.1 bug situation also. In addition, I see annoying bugs changing status to “feature request” for a future version. That of course affects the bug stats, but it does not actually satisfy users.

In related news, Peter writes:

MySQL booth was not to active ether – in many cases I could see lonely Kelly sitting on it. This could be related to the fact there were few MySQL Developers hanging around it most of the time, compared to army of technical guys at PostgreSQL booth. I would say even SolidDB booth was more attended.

Now, why would an OSCON visitor hang out at a booth where only a non-technical marketing person (very friendly, no issue there) is present? Please allow me to introduce to you the left foot of MySQL Marketing. You may notice the gaping hole in it, caused by an unfortunate shooting incident. Fortunately, it now matches up with the right foot.

Posted on 5 Comments

5 thoughts on “OSCON MySQL foo

  1. Arjen
    thanks for your feedback; always appreciated. We have been steadily fixing quite a few remaining issues in 5.0, but I understand your point.

    If you have some suggestions on the marketing front, please don’t hesitate to let me know.

    –Zack

  2. Thanks Zack. I have two suggestions on this front which actually combine.

    1) Seen in a very broad context, Community could indeed be regarded as part of marketing. However, it’s not quite the same as for instance regular product or brand marketing, and some of the approach is fundamentally different. Seeing MySQL Community people do things that are very definitely in the realm of the (narrow context) marketing is IMHO just bad.

    2) There is traditional marketing, but MySQL AB is not quite a traditional company and definitely does not operate in a traditional market place. It is, for instance, impossible to fully segment different markets with targeted approaches, since the materials going out over the net are “out there” and will be publically discussed. Also, (potential) customers are smarter and will actively engage with eachother as well as with other sources of information. They are not merely isolated consumers of a limited set of information.

    People have coined the phrase “Marketing 2.0”. Others have explained that concept better than I can, but in any case the above few examples are part of it. Unfortunately, I don’t see MySQL AB employing Marketing 2.0 techniques. I think it used to, and it doesn’t any more. I currently believe MySQL AB is firmly engaged in Marketing 1.x while trying to interact with Market 2.0. It’s not compatible.

    I reckon that a Marketing 2.0 expert looks a lot more like a business savvy Community person, than like a Marketing 1.0 person. That doesn’t mean that a 1.x person can’t upgrade to 2.0, but it needs to be recognised that it’s not at all the same. Given MySQL AB’s current marketing approach as seen publically, the question begs: does MySQL AB’s Marketing employ Marketing 2.0 experts?

    In the end, it is of course the market will provide the answer. I have no doubt MySQL will continue to be successful in a business sense. But how much of this business will MySQL AB be able to catch, and how ubiquitous will MySQL be in its still growing market place?
    The disruptor is already being disrupted. That’s natural, I suppose.

  3. OK, so I’m wondering why there’s a lot of harshness in this post. Don’t *most* x.0 releases of a product have lots of bugs, and the x.1 releases fare better?

    I saw Kaj’s post as annoying because it doesn’t say much. “5.1 is more stable than 5.0”. Well, duh. 5.0 put in a lot of new features, that’s why it’s not 4.2 or 4.3!

    So how is it a bogus claim? It’s definitely factual — and what would you like a community rep to say — “well, 5.0 was a disaster, but 5.1 is better!” That’s spin as well! Everything is spin.

    I don’t see a problem in saying 5.1 is much more stable — particularly from a community that KNOWS 5.0 was a disaster. And Kaj said that the company was “happy” with 5.1 — maybe that implies they weren’t happy with 5.0?

    It sounds like you’re upset that there are many bugs that aren’t fixed, and are using that to attack Kaj’s statement. But I’d like to know — *is* 5.1 more stable? The company has to have an opinion, after all, they’re the ones that determine when a release is alpha or beta or GA. Should the community decide that too? Should the community decide when a release is end-of-life’d too?

    Sorry, but it sounds like you have pent-up bitterness. Of course you know the inner workings better, so you may have the message right but the delivery wrong, or something……but I didn’t particularly see this as a community damaging event.

  4. The MySQL major releases use fairly arbitrary numbering. 4.1 was as significant as 4.0, and 5.1 diddums when compared to 5.0. Just check out the new feature set!
    5.1 puts in a pretty large and fairly involved range (in terms of server impact) of new stuff, like partitioning, disk storage for cluster, and so on.
    4.1 added subqueries (serious changes in the parser, optimizer, query execution), multi-byte character set (major changes in the entire string handling infrastructure)… it was not a minor release by any means.

    I don’t care whether 5.1 is more stable than a continuing disaster. I just care whether it’s usable on its own merits. The flow fom alpha to beta and onward is fairly well defined (it’s in the manual), and when adhered to it’s quite transparent – when looking at the bugs system you’d be able to discern where it’s at.
    GA is a bit of an arbitrary judgement and that’s fine. It needs to be done as some companies have a policy that only releases branded “GA” can be used.

    But in the end, it is indeed the userbase that decides when a product is stable. This is proven by the fact that many companies use MySQL versions in production before GA stage, and other companies don’t use GA versions for quite a while after release. People do their own assessment, based on their needs (and some company politics).

    There’s no need to drag that into the extreme… end-of-life is a business support decision. In principle, community could decide to backport some fixes from a later version and thereby continue the life of a release. This is in fact what Debian has done for a long time.

    Regarding spin and delivery. Kaj heads, as you know, the Community department at MySQL AB. He is not in a PR department, nor is he part of marketing. Producing spin is specifically not part of his job description. The Community department is separate from Marketing, Sales and the other departments for good reason. Their tasks and methods are (or should be) different. If not different, what’s the purpose of their distinct existence?
    I’d be just as annoyed with people at other companies doing the same. A few years back now, Zend were kinda plugging themselves as the inventors and owners of PHP, which was of course a major spin on reality and a rewrite of history – obviously it ticked off many in the PHP community. Fortunately, Zend has ceased that angle, I believe the change came about with different staffing in their marketing department.

    The point was (and is) this: companies and products are cool because *others* say they are. Talking yourself up is utterly uncool. A whitepaper does not make a product any better.
    Credibility and respect comes from your actions and judgement of those actions by the peers amount the users.
    When people who are respected in the MySQL community say “I’ve been using 5.1 and it’s solid”, that’s something I take note of. MySQL is a good product, and so I think it’s unwise for MySQL AB to talk things it up themselves (particularly coming from a community rep), as that merely distracts from, and dilutes, what others might have to say on the matter. Some may not even bother now. A great pity, as these insights from actual users are clearly the most valuable.

  5. that is realy nice article, increasing my knowlwgde about what the game Mysql is playing in version issuing like 5.0 and 5.1 thanks for that
    ————–
    Ahsan Shahzad
    Mysql Eaxamples
    http://mysqlexamples.blogspot.com

Comments are closed.