Cloudscape Gains Momentum

A reader writes: "There's been a lot of bits written recently about the growth of Open Source databases; as well as IBM's patent gift, as their release of the Java database Cloudscape. There's a contest running on SourceForge.net around Cloudscape; download and run with it." SF.net is part of OSTG, like Slashdot.

Comments

Java database ?

[Anonymous Coward]

wouldn't it be quicker to print your data on paper and sort it by hand ?

Re:Java database ?

[Eric Giguere]

Hey, I like Java as much as anyone, but if you're looking for a fast, multi-platform, zero-administration database be sure to check out SQL Anywhere Studio. The Developer Edition is completely free and runs on Linux, Solaris, Mac OS X, Netware and most flavors of Windows, including CE/PocketPC. See the description of what's in SQL Anywhere Studio for details. (Note that some of you may remember SQL Anywhere as Watcom SQL.)

Eric
Listen, folks: JavaScript is NOT Java!

Eh..

[oisteink]

And the story is about?

marketing?

Re:Eh..

[Jackdaw Rookery]

I think you may have missed the point of the post, the story is advertising Slashdot jumping the shark.

Well OK maybe not, but this is a fluff piece at best, self advertising at worst. Plus the prize is US only so it cuts out a lot if interest right away.

Nothing to see here, move along.

Wrong language, wrong thing.

[InterruptDescriptorT]

A Java database?

Isn't that a little bit like writing a Fast Fourier Transform in LISP?

There are lots of things that Java is perfectly suited for. Databases are not one of those things.

Re:Wrong language, wrong thing.

[sporty]

Why not? If it's fast enough, and it works, it'd be good for java embeded applications and/or stuff that uses a database for scratch-work. It could also be used instead of complex data structures in memory that require massaging to get anything out of.

Re:Wrong language, wrong thing.

[sonofagunn]

Have fun embedding your C database inside a Java server app (I think you've missed the point of Cloudscape).

Re:Wrong language, wrong thing.

[EkkiEkkiShiwaddle]

There are lots of things that Java is perfectly suited for. Databases are not one of those things.

I'm a Java developer, and have been almost from the very start of Java. While I tend to agree with you at first, I would like to point out some benefits of using Cloudscape.

Over the years I've had a few (small, hobby) projects, where I really felt the need for a database which did not need to be fast nor did it need to be huge. So I tried Cloudscape. And while it isn't the fastest around, it gets the job done.

The easiest part is that you can just bundle it with your application, without writing a single line of code!

I've tried MySQL, but I still need to install it prior to being able to run my application - no need for that with Cloudscape.

But, granted, it is not the fastest out there. If I need a *real* database, I just use Oracle.

Re:Wrong language, wrong thing.

[MarkWatson]

Common Lisp is fast and efficient for numeric calculations (*) - I have used it for large neural network training runs, FFTs (I found a library -did not write it myself), etc.

While Java and Python are the languages that I most frequently use, CL is a powerful tool - try it!

(*) with compiler type directives

Re:Wrong language, wrong thing.

[Empty Threats]

Trolling about Java and Lisp at the same time? Isn't that a little bit like ASKING FOR IT? Fast Fourier Transform is actually one of the traditional benchmarks for Common Lisp, the Gabriel series. Now run along and play in traffic.

Re:Wrong language, wrong thing.

[s1234d]

Actually, modern Lisp implementations can be faster than compiled C. Some Lisps run like lightning on numerical stuff. See here for an example:
http://home.comcast.net/~bc19191/blog/040308.html

As usual

[cpct0]

Only one entry per person and you must be a US or Canadian (except for Quebec) resident over the age of 18 to play.

I'm from Quebec ... and as usual, contents makers will not take the trouble of being compatible with Loto-Québec's rules. They are easy to follow, strangely... not a lot of stuff to fill in. It mainly is there to make sure we don't get screwed up

... and it also means everyone in the other countries will not be able to participate. Yeehaa

What's in a name?

[Trejkaz]

Wasn't Cloudscape donated to the Apache project, where it changed its name to Derby? Don't IBM think that offering a download of "Cloudscape" is going to confuse developers?

Spelling mistake in submission

[Anonymous Coward]

The submitter mispelt 'cynical ploy' as 'patent gift', wake up editors.

Re:Spelling mistake in submission

[mabinogi]

no, he misspelt "misspelt".

I've written a review of this

[Simon Brooke]

As it happens I've already written a review of cloudscape; the google cache of it is here.

Brief summary: get the Apache version; reasonably full SQL92 syntax; performance OK; a bit lacking on security.

Re:I've written a review of this

[inertia187]

reasonably full SQL92 syntax;

Actually, I think the biggest limitation is the 18 character names for constraints (et. al).

Re:I've written a review of this

[Oliver Aaltonen]

You probably want to use the text-only cached link, since the one you gave still grabs files from your web server (images, stylesheets, etc.).

Re:IF......

[Anonymous Cowherd X]

You can win an Ipod or TShirt only if you are US or Canada resident....are they trying to encorage inmigration? The rest of the world should not bother about using Cloudscape I assume...

Let's see, most people would need at least an hour to write a JDBC application to search the tables for some silly magic word. Most eligible contestants from the US and Canada who have the knowledge to write such an application make more than enough in one hour to buy an iPod and a T-shirt. So why should anyone even bother to play this game, just for the fun of it? You develop JDBC applications all day long and then in your spare time you do the same for fun... and you might even win a T-shirt. Wow, they really know how to motivate people! Alienating the rest of the world and giving people a sense of worthlessness about their achievements is not the best way to make a product popular. But I'm just clouding the issue...

Here is a swell idea!

[Anonymous Coward]

Lets see how many open source projects we can get going at one time to compete against one another with absolutely no standardization! YESSSS!

But seriously -- this is getting out of hand. If every single company, individual, etc. starts an open source project what does that accomplish? Nothing. It brings about 100 different alternatives to the market most of which would never be viable in a commercial environment.

News flash: devote time and effort to succesful projects that extend the reach of the open source community. Another open source database project is just well... another project.

Re:Here is a swell idea!

[karakal]

If you, my friend, didn't get it: this is the idea of OS. Not to be restricted to some few projects. And I always find something with stands for or against any new project, when I am devoting some time to it.

Re:Java IS slow

[WillerZ]

Portable Java is slow, because you have to use the abstract types sun give you rather than exploit your platform's capabilities.

The JVM initialisation time is waaay longer than the C runtime initialisation time on every platform I've tested, which makes short-running processes feel awful in Java.

Finally, whenever it becomes necessary to actually _do_ anything, the JNI must be crossed, and that's slow.

Phil

C++ Databases and Open Source

[plasticmillion]

I was struck during our search for a database for our product by the paucity of systems that fulfill our requirements:

1. Small and embeddable
2. Open source
3. Native C++ interface

It seems like practically all the databases that fill the first two requirements are written in Java and use it as their primary API. Finally we settled on Sleepycat's Berkeley DB XML, which meets these requirements and has a native XML interface to boot. The one potential caveat is that they use a hybrid license that lets users "opt out" of the GPL by paying a fee, after which they can distribute their product as closed source. Personally I think this is a great thing (always nice to have more options and you don't have to pay anything if you product is also open source), but I'd be interested in the views of the /. crowd.

Re:C++ Databases and Open Source

[defMan]

How about using SQLite. It's a C library (native C++, does that mean C is ok?) which provides a self-contained database engine. No need to run a DBMS.

Re:C++ Databases and Open Source

[Pedersen]

You might wanna check out SQLite, which is in the public domain. It matches all of your requirements, and then some, I think.

Purpose

[barryman_5000]

Best thing about cloudscape is small embedded databases for java programs rather than making calls to a huge mysql database. You may say that java is slow but anyone using cloudscape for more than 20,000 entries may not be thinking that through. Cloudscape will also be useful for java programmers to program a database using java.

2 points you are wrong on.

[tod_miller]

Write a program in cpp. Run it. a million times.

Now write a program in Java. Guess what, Java is not an interpretted language. It doesn't use some fluffy brightly coloured play-doh type bits, it uses regular ones, like cpp binaries.

Java compiled the byte code into executable form. Then it does some on the fly optimisation.

Java VM's can out-perform cpp over n cycles, where n is sufficient to overcome Java start times (seconds)

Secondly, FFT or a DCT in Java or cpp, I haven't seen benchmarks, but there are right and wrong ways to do it in both languages.

Thirdly, Databases, noone said they had to be FAST. Now remember Java isn't slow. Most databases on this earth spend thier time knitting and throwing stupid errors.

A few have to deal with mind-GOOGLING amount of data.

So there you have it, if you want a very nice Java db, try mckoi

http://mckoi.com/database

great support, open source, and works like a charm! I use an IBM DB server remotely, and an embedded mckoi instance for such things as prefs and other niceties. (as well as local work on data)

There you go.

Re:2 points you are wrong on.

[Anonymous Coward]

Java VM's can out-perform cpp over n cycles, where n is sufficient to overcome Java start times (seconds)

Yes, and a well-supplied tortoise can out-distance a rocket over n years, where n is much longer than the rocket's burn time :/

But that is not embeddable

[MarkEst1973]

Your post was certainly informative in that others might have learned about a new tool and are looking at Sql Anywhere, but to be fair, the purpose of that tool is different than Cloudscape's.

Embedding a database in an application can be very useful, such as in a desktop GUI where you cannot rely on network communication or maybe don't want to bother with a client/server environment.

I'd certainly consider Java/Cloudscape for a desktop db-backed application over anything built in Access.

As always, fit the tool to the job, not the job to the tool.

Re:But that is not embeddable

[mod_critical]

I have found the ability to embed the database in the application exceptionally useful for a recent J2EE deployment we worked on.

I liked the idea of just being able to plop the org.apache.derby.jdbc.EmbeddedXADataSource class into my connection pool and have a database. I like that I can tell the suits that the database is built by IBM and that's all I hear about it. And I like that once this venture grows and starts to make more money I can switch fairly seemlessly to DB2.

It probably should not be considered a security measure, but there is also a certain level of security gained since there is no actualy database server that can be accessed seperately from the application server

Re:But that is not embeddable

[duffer_01]

"Embedding a database in an application can be very useful, such as in a desktop GUI where you cannot rely on network communication or maybe don't want to bother with a client/server environment."

I completely disagree. I think most people think that SQL Anywhere is just used for client/server communication when in fact you can also easily use it for standalone use in embedded applications. In fact, I would argue that SQL Anywhere is far easier to embed than Cloudscape since you only need to deal with 2 database files (db and log) which can be copied from any of the supported OS's (Linux, Win32, WinCE, etc). You can simply embed the dll's and exe's that you require for you application directly into your install. Add the fact that SQL Anywhere takes a lot less space than Cloudscape and you have a much better embedded solution.

Arrrrghh!

[Moe Yerca]

You bastards! Thanks for cutting my chances of winning an iPod down 100x.

So how does it compare to the alternatives?

[Anonymous Coward]

Namely HSQLDB and Mckoi?

Anyone find a decent article comparing the three yet?

It is fast

[sasoon]

Download and you will see that it is fast. Have not tried Derby, but I tried http://mckoi.com/database/ 2 years ago. It was very fast. Under win32 I compared it with a MS jet engine database (mdb file) with the jdbc odbc bridge driver, almost equal speed, plus the possibility to work with unicode strings.

Re:Java zealot's response

[Ryosen]

Oooh, lookie! It's a trolling AC. Ok, I'll bite.

Java *is* a mature language celebrating its 10th year in 2005. J2EE, as the sum of its collective parts (EJB, Servlets, JSP, RMI) has been around for 7 years. By industry standards, these are "old-timers".

Java is memory "conservative" (on a server), requiring far less RAM than the .NET platform. Depending on the application server that you implement, you will find varying degrees of performance. Try not to give in to the uninformed mindset that the Java runtime included with the SDK in 1996 was the end-all be-all of Java compilers. That FUD went out of style a long time ago along with bell bottoms, the Dodo and civil liberties.

Java *is* platform-independent. My firm has developed applications running on various different *nix and Windows platforms that have been in production for over 4 years. Any platform-dependent implementation (e.g. via JNI) should be abstracted far enough away from the core application that, in the rare case that it is required, it can be easily replaced. To not design your application in such a manner is foolish and irresponsible.

Business has been very successful on the Java platform for years and much ealier than .NET and ASP.

Or are you suggesting that everything would be better off if it were just writtin in Perl?

Silly troll.

Re:Yes, it is.

[Ryosen]

I think you might be misunderstanding what JNI is. Each Java runtime provides its own implementation to the native underlying system libraries. Java does not communicate through JNI for I/O of any kind.

JNI itself is a generic abstraction layer to the underlying operating system. It provides a mechanism whereby dynamic libraries not directly supported by the runtime engine can still be accessed by a Java application.

Threading, sockets and GUI are implemented via the native system libraries. AWT used to be very slow (as opposed to just being merely slow today) due to its own multi-layed abstraction. But at no time was JNI the conduit for these systems.

Further reading: http://java.sun.com/docs/books/tutorial/native1.1/ concepts/index.html

Cloudscape is great for automated unit testing

[MagicMike]

I was having a hard time maintaining a distinction between unit testing and integration testing with some of my back-end code until Cloudscape came out.

Now, instead of needing to have a container up, or have a database running or similar, I can embed cloudscape (which is quite small, really), set up a group of in-memory tables with the state I need, execute a unit, then check the tables for correctness and destroy the database.

Shazam, database unit tests - no external dependencies.

A lot of people are mumbling and complaining about Java and databases and speed, but apparently they've missed the whole Moore's law thing. I'm betting they either haven't tried it or they have some over-inflated view of how fast their applications need to be (or how much it costs to get fast hardware vs. their time).

This stuff easily meets the "fast-enough" test for lots of uses, unit testing being just one of them.

Apache at the ASF

[dirkx]

Yes it was - see the Derby project in the Incubator.

Beeing in the incubator does not yet mean it is a full blown ASF project - but that the ASF is in the process of ensuring that there is a healthy community around it, that all the legal paperwork, trademakrs, grants, copyrights and other interlectual rights are sorted out, that commiter license agreements are on file for each developer, etc, etc..

Once that is all in place (and getting a healthy long term community is hardest - the rest is just endless grunt work and digging through code and legal paperwork dotting i's and crossing t's) it'll leave the incubator and be a full blown process.

Feel invited to join and make this happen ;-)

Dw.