Comments on: Back to the future for MySQL

MySQL hasn't always been open source. And it's not clear that proprietary software will help it any better now than it did in the past.

[seo2seo]

The point you did not pick up is that Open Source relies on the freely given contributions of many people, though the big stuff may be done by a paid core team. Once you start to sell your product (or re-start), then you have a choice; you either reminuerate your volunteer helpers, or you don't.

And in the real world, you don't.

And that kills the open part of open source. They don't want the money, they did it for love. But neither do they want you to get rich off their backs.

[seo2seo]

The point you did not pick up is that Open Source relies on the freely given contributions of many people, though the big stuff may be done by a paid core team. Once you start to sell your product (or re-start), then you have a choice; you either reminuerate your volunteer helpers, or you don't.

And in the real world, you don't.

And that kills the open part of open source. They don't want the money, they did it for love. But neither do they want you to get rich off their backs.

[ldubost]

"Is it fair or productive to close off the code after open source has made it popular?"

There are in reality two questions here:

1/ Is it fair

For sure there is an issue here. An open source company is at the same time a business and it's license (double license) does not guarantee that they will continue in the same business model. Contributors and others have helped the software improve and spread and now they see a slowly but gradual divergence from the initial path they decided to support.

You can rely on "human decision" or on "market competition" to make sure they do stay in the same business model.

Now in the case of MySQL, You say Marten and folks mean well and usually do well. Let's face it. They have cashed out and the boss is now Sun. And Sun payed a hefty price and needs ROI.

When it comes to "market competition" I think it's the same as "is it productive".

2/ Is it productive

If you go down the road of closing, other open source products might get more traction if the strategy changes. Customers might feel played and you loose sales productivity. The double license might turn out in a pure GPL fork by the community. But saying that market competition will be sufficient to stop the trend to higher ROI through closing would be like saying that proprietary software companies are dead because of the open source competition. I don't believe this is true, although they will definitively see their revenues and margins crunch, but in my view it is at least like the Mainframe market. That's not were the innovation lies but you can still get tons of cash by servicing your customer base.

This is the same for MySQL, they can decide to serve their customer base. It has good changes of being short term productive by making more short term cash, but this also leads to have the open source powered viral marketing stop.

3/ But what is the big question ?

The big question is actually: "When open source means investors and the stock market, who gets the final word ?"

- is it the fairness ?
- is it the market ?
- is it the investors ?
- is it the founders ?
- is it the open source license ?

For sure we could dream of an ethical world where it's the fairness ! We could dream that the founders that have cashed out will still have a word to say. Of course the ones that finally decide are the owners of the company and the management team that these owners have put in charge. But they will usually have to follow a few rules: "the market" and the "legality". So we end up with the final sayers being:

- is it productive in a competitive market place to close the open software that made you known worldwide ? And that we seem to agree that we don't really know and that the market will show it. In any case when the open source company decides to do it, the community has lost something.

- is it legal to turn an open software gradually to close source ? For me this ends up being the big question. And although it's not fair to the contributors, the GPL Double License business model makes it legal.

What's really important here is that other open source licenses won't allow this, like the GPL and LGPL without double license. Also a commitment to open source and openly downloadable products, like the OpenBravo commitment ( http://www.cnet.com/8301-13505_1-9920203-16.html ), would help (although it's not sure how binding a commitment like this is).

This is the reason why for XWiki (I'm the initial creator of XWiki both the software and the company), I choose the LGPL license without a double license constraint. This license gives sufficient rights for users (including commercial usage and integration possibilities in proprietary products) and it also binds the users (including XWiki, the company) to continue publishing under the LGPL open source license.

This is binding and will guarantee the continuity of the open source commitment, even if the ROI, the investors or the founders say otherwise. What's still required, is that the company can make sufficient money to fund its share of the development.

We at XWiki will are looking into writing a similar commitment because we think it's important.

I've also published this comment as a blog article at http://www.xwiki.com/xwiki/bin/view/Blog/OpenSourceCommitment

[ian.waring]

One of the gotchas of open source is that a lot of the software "just works". With this, it is easy for a user to slip into a typical "licensed" software model of paying a fee once (in the open source case, for the first annual subscription), and to find that there's little perceived value in telephone support or updates from year 2 onwards - as their experience shows that the software works and support services are very under utilised.

Red Hat used to ask people to activate their subscriptions at time of installation, and found that only 30% or so did so. Imagine that - only 30% of customers buying support actually used it. Hence changes they made over a year ago to activate every user at time of sale, and to change their terms/conditions also to say that if you ran Red Hat on a server, that it was compulsory to have a subscription active on that machine. So, Red Hat have deviated quite widely from "deploy everywhere, only pay us if you need support" that used to be our sales go-to-market in the early days.

MySQL have been offering custom licensed versions of some of their products for some time; my own company have customers who buy MySQL Cluster (where a database sits in shared memory between machines) and appear to be happy to pay into 6 figures for this (it's used in VOIP and Mobile switching platforms). While most customers are happy to pay for a MySQL subscription, the challenge has been to simply communicate part numbers and pricing for them to do so - and to keep it simple. MySQL users pay per server and with no user CALs - at price levels which are tremendously good value compared to virtually every other DB vendor.

The potential for MySQL to derive more value - I believe their community:commercial or unpaid:paid subscription ratio is something like 1000:1 - is without question. I think an awful lot of companies using their software would be happy to pay for the value they receive from their use of MySQL's products as the process to do so becomes much easier. We're already seeing this starting to happen.

Ian W.

[cbbrowne]

"The point you did not pick up is that Open Source relies on the freely given contributions of many people" - is not true in this case. MySQL AB does not depend on contributed code - their software is entirely proprietary, which is an essential prerequisite to being able to "dual license" it. MySQL's software is written as a "work for hire," which undermines claims about it being a "labor of love." The two are not forcibly mutually exclusive, but it IS certain that it's "work for hire."