Slashdot Log In
Sun Announces New MySQL, Michael Widenius Forks
Posted by
CmdrTaco
on Thu Apr 23, 2009 08:31 AM
from the this-cracks-me-up dept.
from the this-cracks-me-up dept.
viktor.91 writes "Sun Microsystems announced three new MySQL products: MySQL 5.4, MySQL Cluster 7.0 and MySQL Enterprise Partner Program for 'Remote DBA' service providers."
which showed up in the firehose today next to Glyn Moody's submission where he writes "Michael Widenius, founder and original developer of MySQL, says that most of the leading coders for that project have either left Sun or will be leaving in the wake of Oracle's takeover. To ensure MySQL's survival, he wants to fork from the official version — using his company Monty Program Ab to create what he calls a MySQL "Fedora" project. This raises the larger question of who really owns a commercial open software application: the corporate copyright holders, or the community?"
Related Stories
[+]
Has MySQL Forked Beyond Repair? 334 comments
snydeq writes "Fatal Exception's Neil McAllister questions the effect recent developments in the MySQL community will have on MySQL's future in the wake of Oracle's acquisition of Sun. Even before Oracle announced its buyout, there were signs of strain within the MySQL community, with key MySQL employees exiting and forks of the MySQL codebase arising, including Widenius' MariaDB. Now Widenius' Oracle-less Open Database Alliance adds further doubt as to which branch of MySQL will be considered 'official' going forward. 'Forks are a fact of life in the open source community, and arguably an entirely healthy one,' McAllister writes. 'Oracle just better hope it doesn't end up on the wrong side of the fork.' To do so, he suggests Oracle will have to regain the the trust and support of the MySQL community — in other words, 'stop acting like Oracle.'"
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
It depends (Score:5, Insightful)
It depends on the license of the software. Always.
Re:It depends (Score:5, Funny)
It depends on the license of the software. Always.
It also depends on where you live. In Soviet Russia, software owns you.
Parent
Re:It depends (Score:5, Insightful)
Well, even if the maintainers have the copyrights, that only means future versions can be closed source. They can't terminate the already-outstanding licenses without a breach of terms. They also own the trademarks to the MySQL also.
IMO, Sun lost the hearts and minds of the developers which is where the real value was. The trademarks and copyrights are worthless if the community views MySQL's direction is wrong and moves entirely to a fork.
Parent
Re:It depends (Score:5, Insightful)
On the flip side, the forking company can't use the same business model as MySQL AB. Since MySQL owned the copyrights, they could see non-GPLed versions of the software under terms that were more palatable to corporations. To a certain degree, it served their purposes to fuel GPL fears.
Now that the forking company is 100% bound by the GPL, they must attempt to undo any misplaced fears about the GPL and seek to convince companies that what they really want is a support licene, additional tools, or trained consultants.
Parent
Re:It depends (Score:4, Insightful)
Most companies don't need a more "palatable" license for an RDBMS. They typically
use it as a product, not something to build a product from. This is the key area
where the GPL can be a problem for a corporate entity. Most of Oracle's database
(or apps) customers don't have any reason to be concerned about their RDBMS having
a copyleft license.
They want assurance that their data will be protected and their operations won't suffer outtages.
Parent
Re:It depends (Score:5, Insightful)
Parent
Re:It depends (Score:5, Informative)
What do you think most companies do with an RDBMS if they are not building software on top of it?
Most people don't build software on top of an rdbms, they build software that uses and rdbms as a backend data store.
Parent
Re:It depends (Score:5, Interesting)
And here is the crux of the argument and why MySQL is doomed...
You see if you have a product that binds to MySQL you will have to GPL your product. Why? Simple...
1) All (most?) drivers are GPL'd.
2) The MySQL notation uses a specific parameter delimintator that is specific to MySQL. And a 4 year old court decision said that there is no binding between application and RDMS if the same code can be used on other databases. With the special notation, it is not possible and hence constitutes a GPL binding.
Personally I see MySQL falling off to the way side...
Parent
Re:It depends (Score:4, Insightful)
Or even better yet. Don't use MySQL.
Parent
Re:It depends (Score:4, Informative)
The case related to a company who used ODBC and whether or not they binded to MySQL. It was not the NuSphere case, but one that used ODBC and MySQL.
The question was if your application used ODBC and MySQL was it binding in the GPL sense?
The answer was in the fact whether or not the application could function with another database. At the time the result was that MySQL lost the case since the application could function with another database.
It was around that time MySQL GPL'd all drivers, and changed their syntax so that it would only work on their servers. That way it is a GPL binding as per the court case.
Parent
Re:It depends (Score:4, Insightful)
if you know the protocol, you can interface with oracle, mySQL, sybase, whatever without touching oracle's code, which means oracle's license is irrelevant if all you want is to build a client app.
Until you want to sell the client app.
Parent
Re:It depends (Score:5, Interesting)
Sounds like you haven't been following MySQL AB very closely. Their interpretation of the license was that any time you paired a MySQL database with an application, you needed a MySQL commercial license. Only if the application supported but was independent of MySQL would you not need to follow the terms of the license.
MySQL even tried to reinforce the idea by purchasing all the third party drivers and changing the licenses to GPL instead of LGPL or otherwise.
While MySQL's licensing info has changed over the years (interestingly not archived by the WayBack Machine...) even their current page on licensing [mysql.com] is designed to steer users toward purchasing a commercial license:
For quite a few legal departments I've worked with, "the GPL does not apply" is magic words to their ears. They will instruct the business to grab the commercial license to get around the restrictions. In addition, there is the MySQL libraries issue I referred to above:
The MySQL forking company is going to have to undo all of the anti-GPL ideas they've been riding, and convince companies that they don't need a commercial license. (Since it's not in the forking company's power to provide one.)
Parent
Re:It depends (Score:5, Insightful)
Agreed. See x.org for how quickly a community can switch to a fork.
Parent
Re:It depends (Score:4, Informative)
And also consider that Oracle also has the Sleepycat Berkeley DB engine, which you really have to know where to find to get.
And they also have the old Digital database engine.
Their method is to acquire competitors and then slowly decrease development and avoid promoting the products. Then the products can silently die.
But I wonder if they haven't bitten into something a bit too hard to bite into this time...
Parent
Re:It depends (Score:4, Informative)
Oh, sorry to reply yet again. Yes, I can name a number of project forks in which the original fell by the wayside as a fork took off. Not all the originals are dead, mind you, but the forks are much more popular. A few of these are situations in which the original is still viable (Debian, for example), but in which the fork has a huge number advantage or a lot of momentum.
GCC -> EGCS -> GCC
Mosaic -> Netscape
Netscape -> Mozilla
Mozilla (Seamonkey) -> Mozilla Firefox
KHTML -> WebKit
Debian -> Ubuntu
XFree -> X.org
StarOffice -> OpenOffice
SSH -> OpenSSH
Hack -> NetHack
osCommerce -> ZenCart
AT&T Unix -> BSD
Parent
Who Owns Open Source Applications? (Score:4, Insightful)
This raises the larger question of who really owns a commercial open software application: the corporate copyright holders, or the community?
No one. Or, perhaps, everyone. That's kind of the point, isn't it? It isn't locked into anyone's individual grip.
Cases Like OpenOffice.org (Score:5, Insightful)
This raises the larger question of who really owns a commercial open software application: the corporate copyright holders, or the community?
No one. Or, perhaps, everyone. That's kind of the point, isn't it? It isn't locked into anyone's individual grip.
"Open source" is just too broad a term to address this way. You would have to look at individual licenses. On top of that, you have things like Open Office, which is "open source" but clearly controlled by Sun (or Oracle now I guess) [slashdot.org].
While you claim you can always fork an open source project, it's not always that simple. Especially in massive open source efforts (like Linux) where they have contacts and knowledge that are vital to the project. It isn't possession or control or fiscal ownership but instead a name you've made for yourself as the Father of some project that gives you "ownership" or "rights." And usually the market share of your user base reflects that.
You'd be surprised how many of your open source solutions are actually controlled and operated by a single entity. And this is great for those products because the entity is usually donating a lot of time and money to it. Should the entity ever drop out, that's when someone can pick up the cross and take it a new direction with everyone helping.
Parent
Re: (Score:3, Insightful)
No one. Or, perhaps, everyone. That's kind of the point, isn't it? It isn't locked into anyone's individual grip.
In case of mysql I think they made a living on selling versions without the gpl license. That business model will not work for a gpl fork.
It's GPL. (Score:5, Funny)
So the answer is yes.
Licenses (Score:3, Insightful)
To ensure MySQL's survival, he wants to fork from the official version -- using his company Monty Program Ab to create what he calls a MySQL "Fedora" project. This raises the larger question of who really owns a commercial open software application: the corporate copyright holders, or the community?"
That's what all the lawyering over the license text is all about. This question is one of the more settled questions in the industry.
Get it here (Score:5, Funny)
Re:Get it here (Score:4, Interesting)
Skype is one, if not the biggest users of PostgreSQL and has released some tools that they use to manage their PostgreSQL cluster [skype.com] that are probably worth looking into.
With postgresql, you also have the option of scaling vertically instead of just horizontally. It seems that Postgresql scales better than MySQL across multiple cpus/cores [tweakers.net] and handles heavy load better.
Another interesting benchmark shows that both Postgresql and MySQL handle load better on Solaris instead of Linux [tweakers.net].
Parent
I think it's time to switch... (Score:5, Insightful)
...to PostgreSQL. Seriously, I already use it for GpsDrive. Now I just need to convince the Cacti devs to switch over.
I'm so going to get flamed... (Score:5, Interesting)
As the owner of a software development company I think your would have to be stark raving nuts to open source your main product. It's not that the model can't work it just that if it becomes successful you are pretty much guaranteed to lose control of it at some point.
If we look at MySQL for example: here's a company that produces half way decent database engine that that make open source. They play the open source game "properly" producing code that a mortal can compile to get a working database. While the company is giving the community what they want everything is hunky dory and there is peace.
Enter Sun who buy MySQL and suddenly the community isn't happy and it's fork fork fork. Only one of those forks needs to be any good and all of a sudden Suns not bought very much at all. If a company plays nice with the open source community forks are fairly easy but rare. The problem is they hang like a knife (or maybe that should be fork) over the company and if they are unfortunate enough to annoy the community they could eaisly lose control of their product.
That said I think there are situations where companies can participate in open source. The Linux kernel and Plone being a couple of good examples. Both of those projects are structured very differently to the MySQL situation though as no one company is trying to make a living off the code.
Re:I'm so going to get flamed... (Score:5, Insightful)
Congratulations! With your very example you actually managed to disprove your original assertion. See, your original claim was this:
"As the owner of a software development company I think your would have to be stark raving nuts to open source your main product."
But, the very first paragraph in that quoted text demonstrates that isn't actually the case. The community was very happy with and supportive of MySQL corporate.
The problem, as you pointed out, was the purchase by Sun. In that case, the customers didn't feel Sun would necessarily have their interests at heart, and so there was dissatisfaction. This is only increased by the fact that Sun has now been purchased by Oracle, a company that actually markets a product in the same general space (I would argue they aren't actually in the same market, and so MySQL has little to fear, but... people aren't exactly rational).
So the key to running a company on an open codebase seems simple: keep your customers happy, and don't give them the impression that their interested are being threatened. But, of course, that's a good general rule to follow regardless of the license your code falls under. The only real difference between open and closed source, in this case, is that if the source is closed, you may have achieved vendor lock-in, which gives you more freedom to buttfuck your customers, as they won't have a clear avenue for recourse... but if that's your strategy, well, frankly, fuck you.
Parent
Re: (Score:3, Interesting)
The Linux kernel and Plone being a couple of good examples. Both of those projects are structured very differently to the MySQL situation though as no one company is trying to make a living off the code.
RedHat and SUSE might disagree with you on that one.
Let me make it easy (Score:5, Informative)
If MySQL had a BSD license it would be owned by the community.
If MySQL had a "non-free" commecial license it would be owned by Oracle.
The mess MySQL, and you, find yourselves in is because of MySQL's stupid dual-level license bullshit. Nobody seems to be able to figure it out or agree on it and it has caused more column inches of claptrap on Slashdot than the MySQL/PostgreSQL threads themselves. MySQL's originator's wanted to have it both ways: Lots-O-corporate money AND GPL poster child. Well they got their money alright, but to get it they had to pray for a really wealthy, poorly managed corporation to come along and vet their convoluted business plan. That would be Sun.
Now, with a billion dollars spent to "buy" MySQL but a bunch of forks still out there, no company in their right mind is going to invest anything in MySQL because they'll be worried Widenius will just steal the improvements and fork it again. MySQL is pariah, it's poisoned.
If you're running any kind of data volume worth talking about you're better off with PostgreSQL. Not only is it faster with *real* queries and more robust, but now it's safer going forward.
Re: (Score:3, Informative)
Except MySQL didn't want their code "owned by the community". That's why they dual licensed it to begin with.
If they wanted it to be "free and open like BSD" they could have simply made the relevant parts LGPL.
This notion of yours that FSF licenses are incompatable with business is just a fantasy.
Oracle wants more people writing SQL (Score:3, Interesting)
So they improved InnoDB to make MySql more attractive to the small folks. If they become as big as eBay and PayPal, they probably will switch to Oracle (;-))
--dave
Re: (Score:3, Insightful)
Why would the fork have to stop supporting InnoDB?
Re:Right (Score:4, Informative)
Did anyone else notice that his little toy database is practically useless without InnoDB, which was written by a third party and is owned by Oracle?
If you mean for transactions.
If you want a really fast free database that supports fulltext indexing, and you don't need transactions, MyISAM in the engine to use.
Parent
Re: (Score:3)
Yeah, you're right. Foreign keys, who needs 'em?
Seriously though, this could be good news for PostgreSQL. Fingers Crossed.
Re:Right (Score:5, Interesting)
Seriously though, this could be good news for PostgreSQL. Fingers Crossed.
AFAIK more and more people are using PostgreSQL. More and more providers are supporting it. Five years from now, it could be a whole different landscape...
Parent
Re:Right (Score:5, Funny)
Yeah, you're right. Foreign keys, who needs 'em
They should even get rid of SQL support altogether and just call it My. Who needs all that complicated elitist crap anyway?
Parent
Re:Right (Score:5, Funny)
Operations like Skype [skype.com]
Parent
Re:Right (Score:4, Informative)
Or you can use SQLite, get more speed, and still have transactions. (Although fulltext indexing does require a loadable extension.)
Parent
Re:Right (Score:5, Informative)
SQlite has supported per-table locking for a while, and I believe it supports per-row locking in some situations. It is not designed for concurrent writes, but it can be great for anything read-heavy workloads. It's certainly not suited for situations where you have a lot of concurrent writes, but for a CMS it can be a very good fit.
If you want full-text indexing, transactions, and lots of concurrent users, PostgreSQL is generally a better bet. MySQL is being squeezed at the bottom by SQLite and at the top by PostgreSQL, and both have less restrictive licenses (public domain and BSD, respectively). I'm amazed that it's survived this long.
Parent
Re:Right (Score:4, Interesting)
I started with MySQL because it had replication. Yeah, I know Postgres has slony or whatever it's called, but I cringe at the though of maintaining disparate pieces of software (that sometimes don't get along) on something as critical as a database. Maybe it's changed, maybe it hasn't, but that's how it was when I chose to use MySQL.
Parent
Re:Right (Score:4, Interesting)
Parent
Re:Right (Score:4, Informative)
The point of MySQL isn't a "lightning fast data indexing/accessing machine". The point of MySQL is the modular backends which enable it to serve as a common gateway to tables that each use the storage engine most appropriate to the way the table is used. (some of which may require a lightning fast data indexing/acessing engine and accept some risks to get it, some tables may not.)
The point of InnoDB (and, presumably, Falcon) is to support the kind of usage scenarios for which traditional RDBMS are designed, while the point of certain other MySQL table drivers is to support other types of loads.
Parent
Re:Should Read: Sun announces last MySQL products (Score:4, Insightful)
I have to disagree, I think that would be immensely stupid of them. I think they'll just use it to try to funnel users butting up against its limits towards full Oracle. If they kill it they lose that potential sales channel.
Parent
Re:Should Read: Sun announces last MySQL products (Score:4, Insightful)
Who said anything about closing the gap? Continuing to develop and support MySQL doesn't mean turning it into a powerhouse database like Oracle.
The simple fact is, MySQL and Oracle do not, and have never, played in the same league, and I believe it would be a mistake to try and turn MySQL into a shitty Oracle. MySQL has a niche... keep it there.
Parent
Re:Should Read: Sun announces last MySQL products (Score:5, Interesting)
The gap between MySQL and Oracle is huge and not likely to be closed anytime soon.
Technology leaders in big companies aren't as into all the open source gossip as the slashdot crowd are and I wouldn't be surprised if many of them didn't even know there were MySQL forks or what that meant.
They would rather go with a MySQL that is named MySQL and has a big company like Sun or Oracle, the leading db vendor that also owns the only sane database engine for MySQL, than some noname fork. Even if it was started by the MySQL founders and all the developers went to it. If all the MySQL developers go to a fork, well then Oracle developers will take over.
What's more concerning is IBMs partnership with EnterpriseDB [cnet.com], which is based on PostgreSQL.
If you want an open source database that closes the gap with Oracle, use PostgreSQL.
Sun should have never bought MySQL. Instead they should have put more effort into PostgreSQL. Sun has had some big wins with Solaris and Postgresql [arnnet.com.au] in the past and offer support for it on Solaris.
Must be tough since Oracle is an important part of Sun's business but Oracle has done things that could be considered as stabbing Sun in the back too.
Parent
Re:Repercusions for FOSS licenses (Score:4, Insightful)
But what license was the FICS code under? Was it really "open source"?
There are plenty of licenses that provide for distribution of source but are so restrictive that no one considers them to be "open source".
Parent
Re:Repercusions for FOSS licenses (Score:4, Informative)
FICS has been replaced by chessd: http://chessd.sourceforge.net/index-en.html [sourceforge.net]
Parent
Re: (Score:3, Informative)
Re: (Score:3, Interesting)
They could argue that, but even if that were to be found to be the case, for any program with significant distribution, I'd think the doctrine of "detrimental reliance" would apply.
Re: (Score:3, Informative)
they way MySQL and stuff from the GNU/FSF is set up, they require contributors to sign the work over to them. Then the body in charge has ALL the rights and can do what they will. Like when the FSF moved everything to GPL 3, they could do that unilaterally because they had assignment.
Mozilla also has assignment and releases just what you say under a tri-license, the same code base published 3 times. Two are open source (MIT & LGPL) but the main Firefox branded binary is actually NOT open source. Anybo
Re: (Score:3, Insightful)
This is taken care of by section 6 of the GPL v2 (though it appears as section 7 in the MySQL documentation for version 5.0 at least.
Re:I Forked a Couple Nights Ago (Score:5, Funny)
The problem with forking is all the child processes, though.
Parent