Slashdot Mirror
Oracle To Offer A Free Database
An anonymous reader writes "ZDNet News reports that Oracle is likely to announce a free version of its Oracle 10g Database. Oracle Database 10g Express Edition will be free for development and production use, and could even be distributed with other products. What does this mean for the future of MySQL and PostgreSQL?" From the article: "By introducing a free entry-level product, Oracle intends to get more developers and students familiar with its namesake database, Mendelsohn said. Those customers, Oracle hopes, will eventually upgrade to a higher-end version."
I've been meaning to ask this for a while on Slashdot, but how many people here use a SQL database as part of a hobby or for fun?
I currently have PostgeSQL running on my Tiger box. I initially installed it just to experiment with SQL and database normalization, but now I keep my comic book inventory on it. (I know that this is like swatting a fly with a nuclear weapon but I enjoy using PostgreSQL and it is FREE software.)
As for Oracle's announcement, I think that it can be a good thing, provided you are willing to live with their restrictions and only need support for Linux (x86?) and Windows.
Sounds to me like the difference is that you're allowed to use this new edition for (limited) production use as well. Now, I'm sure there are a ton of small shops currently using the free, "non-production" edition for production apps, but of course they're not really supposed to; this gives them a legal route. But I still think the deliberately crippled nature of the product makes it unattractive relative to the open source contenders, in terms that even PHB's will understand: "Boss, if we go with 'free' Oracle, we're going to run into that disk space limit pretty fast, and then we'll have to pay $$$."
The correlation between ignorance of statistics and using "correlation is not causation" as an argument is close to 1.
The problem with Oracle is, it doesn't scale at all. It is meant to do grid computing, but can't really do anything smaller.
How often do you need to use a cluster for your data? If you are a major organization, then you will, but the majority of installations are pretty small. Firewall/website logs. Customer data. And so on.
I have once developed a workshift-tracking application for a company with around 200 employees. A couple of years later, the total data takes 17MB. Why would you use Oracle if MySQL works faster and takes 1% of the resources? A minimal installation of Oracle 10g takes ~800MB of memory, and will take over ten hours to install on a machine with 512MB ram, on the other hand, on my firewall (486, 32MB ram) MySQL can handle Apache logs (only about 200k hits, though) taking a split second for any reasonable query.
Oracle works better for clusters.
MySQL works better for a single machine.
MySQL is a lot faster. Oracle takes distributed processing a lot better.
But uhm, where does a crippled version fit in the picture?
The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
You could download and test Oracle's DBs for free for quite a while - but only now you can use them for free in a production environment. That's clearly aimed at MS' offerings like MSDE and SQL Server Express and not at MySQL or PostgreSQL.
Also Oracle is an enterprise DB and MySQL, PostgreSQL or even MS SQL Server can not be compared to it in that regard. This also means you need trained staff to administer it - forget about just downloading and using it. Tried to get a demo of Oracle's XML Publisher working - I know what I'm talking about ;-).
So if your shop is already using Oracle's DBs this is a nice offer for the occasional small project. But for everyone else, just stay with what you know and love - whether it's MySQL, PostgreSQL, MS SQL Server or something else.
Oracle also has a number of certifications that make it very attractive in certain industries like defense, healthcare, etc. I work in the medical field, and a free 4GB-max "validated" Oracle database is a huge boon to a wide variety of medical ISV's. Given that's it free to redistribute, the OSS/proprietary thing kind of takes a backseat for a lot of these ISVs.
.NET's layered data architecture.
By the same token, I also don't think it's going to drive sales in the way they think it will. Databases are slowly but surely going commodity, at least at the lower end of the market, and this merely reinforces that trend. And along with that, there's an increasingly robust set of tools to obviate the differences between these database for most uses that don't demand extreme peformance, from Hibernate and ORM packages to ADODB and other database-independence layers in PHP to
As a Postgres user, I'm hopeful that Sun's proclaimed interest in Postgres will result in this kind of "validation". However, given Sun's reputedly somewhat lackadaisical commitment to staffing OOo, I'm not holding my breath. With Postgres' extensibility and extremely high-caliber core developer base, I think a strong commitment to validation by Sun could make it a real contender in the medium enterprise space. Validate it, clean up a few features (notably auto-vacuum and passable auto-tuning, maybe some multi-master replication), throw in a simple deployment for ORM or database indirection, and you've effectively moved that commoditization up one layer from the small website developer level.
In the long run, I don't see how this gets Oracle out of the need to transition its core revenue off of its database licenses.
What I've seen a lot of are databases designed by programmers - good programmers too, these guys weren't slouches. They didn't have any formal training in databases or maybe they hadn't paid attention. Programmers tend to approach database design by thinking, "what data structures do I need to get out of this?" Someone more experienced in database design approaches the problem by thinking, "what relationships are here and how do I model them?"
Here's a real-world example. A web-based application that I was hired to extend and maintain included a system for users to exchange the lesson plans they created on the site with other users. In doing this, the users built up something like a buddy list. "These are the people that I often share with." Or you can think of it as being like an address book.
With me so far? OK, the programmer (who was very sharp - probably better at this than I am) approached the database thinking, "what data do I need from this thing," and decided that what he needed was a comma delimited list of userIDs. So he physically stored the buddy list in the database in like a char(500) as a comma delimited list.
That was actually great for what he was doing. He was just showing a user their buddy list. Unfortunately, that isn't normalized. So, there is a question you can ask which can't be answered by the data. That question is, "how many people have userID 50 as one of their buddies?"
See, the correct way of doing this is to have a many-to-many relationship which you implement with a table containing just two columns, userID and BuddyID. So if I'm user 12 and users 13,14, and 15 are my buddies, I have three rows in that table:
UserID BuddyID
12 13
12 14
12 15
Now if I need to ask, "how many people have userID 50 as one of their buddies" I can do select count(*) from X where BuddyID = 50
BTW, I actually fixed this one not by normalizing but with a hack. I appended 0 to the front and back of the buddy list, then I could do select count(*) from user where buddylist like '%,50,%' But hacks aren't how I make my living. I'd prefer to do things the right way. What my boss wanted was, every time you look at this page it shows you who has you set as a buddy. Kind of like what Slashdot does with the "fans" page. If it was normalized, that would be a scan of an indexed column. It would be lighting fast. so fast the page would practically load before you even clicked the link. But doing a "like" on a big char field is slow.
There are I'm sure still other questions that the un-normalized database cannot answer. Also there are problems with deleting users and, the big one, overflowing that char(500). "How many buddies can a person have?" I was asked. "It depends" I said. If all my buddies are low number IDs, I can have a lot. If my buddies are high number IDs, I can have fewer. It's all just huge mess! Of course, it worked according to the original specification though.
I have a lot of respect for the guy who wrote it. And I'm not tooting my own horn either. I've have other people look at databases I've designed and just torn them to pieces. This is just one example where I just happen to know that he did it the wrong way.