Advogato blog for burtonator

9 Feb 2001

Fri, 9 Feb 2001 18:36:50 GMT

OK. I haven't posted in a while. Been busy.

Next week I get to go to the OReilly P2P Conference. Should be REALLY awesome. I will be hacking on code until them. We need to release by Wednesday, the day of the conference.

I can't believe I am about to say this... I think we need another programming language. I usually consider myself VERY conservative in this respect. I see many programming languages with almost NO significant reason for existance.

I can hack anything in 12-15 languages. You name it. I only try to hack on 2 - 3 at a time. Java, Lisp (for Emacs), and XML (if you consider that a language) are my current choices.

Basically it would be a combination of Java and Python.

- would remove bad things from Python.
  - no operator overloading
  - no multiple inheritance
- enumerations
- Java/Python Exceptions
- lambda
- format strings
- fully Open Source with a strong document for the language
specification.
- Java bytecode support (for Jython and Java)
- GNU Java Compiler support so bytecodes can be compiled.
- regexp support from day 0
- better core debug (my JavaCore project)
- Java/Python syntax blend
- maybe the format requirement from Python.  I kind of like
this.  Awesome way to enforce language settings

Of course this won't happen. I am just writing it down so that I can clear my head. Work is too busy right now for me to start such a huge project.

I am going to setup slashcode on my openjava.org site. Hopefully after the P2P conference. The GNU Java effort is really going to make things MUCH better.

Kevin

15 Jan 2001

Mon, 15 Jan 2001 09:13:38 GMT

Big hacking weekend. Spent most of Saturday and Sunday working on OpenPrivacy... in one way or another.

Java is in need of a better component management framework. It is basically Component Oriented Programming so it is similar to XPCOM, COM, EJB, Avalon, etc. There are some Zen like differences that you only start to realize when you do a lot of Java stuff.

The system would be very thin and portable to any framework (Servlet, EJB, etc).

I came up with a basic set of initial requirements. Just soemthing I threw together.

- Talon???

- Talon should be able to use its own Components without causing holes in the system?????

- Cross platform, cross project, cross licensing

- Dual licensed BSD Style/GPL

- Thin factory/component mechanism for managing components.

- Pool/Singleton

- Question... what do I do about the interface??

- Initializes itself on request. Basically just a snap-in library that any application can use.

- Throw Exceptions if the object is not available

- Method to deliver properties to a component.

- XML component/object deployment mechanism.

- Serve up Component objects

- CompomentManager

- getInstance() releaseInstance( Component comp )

- If the object is a Singleton releaseInstance is redundant. It is necessary if it is a pool

- Support the ability to obtain an object by a Handle... Example... files, URLs, etc.

- Ship with some basic interfaces.

- Logger (Singleton) - BasicTextLogger - Log4JavaLoggerImpl

- PropertyManager - Implementations: - PropertyFileManager... properties backed by a file.

- ObjectPool

// get the number of objects available for use. - getFreeInstanceCount()

// total number of objects being used elsewhere... - getUsedInstanceCount()

// get the maximum number of objects to serve. - getMaxInstanceCount()

- Need to have a ComponentHandle object so that we have tight bindings between properties and handles.

- ComponentNotAvailableException if the Factory can not serve this.

.......

Also spent more time thinking about how we are going to handle input into a Reputation Server... asumming it is a black box. Advogato of course is one of the examples. It would be a Reputation Calculation Engine and its integration within the system would be provided by Sierra.

Sierra is coming along nicely. Hopefully we will be able to officially announce something soon.

Also spent a few hours moving Arrowhead over into the Sourceforge CVS. Ant Unit Tests are now in Sierra and everything is running nice. Still have a lot of features I want to move over from Jetspeed though.... but no time.

12 Jan 2001

Fri, 12 Jan 2001 07:35:10 GMT

Just sent this to Simpson Garfinkel in response to:

Java: Slow, ugly and irrelevant

Simson.

Just read your Salon article at:

http://www.salon.com/tech/col/garf/2001/01/08/bad_java/index.html

First.. I agree with you.. but this isn't Java's fault. This is SUN Microsystem's fault. They have systematically lied to the developer community, used their influence to pull the wool over the eyes the press, and basically do everything possible to ensure that Java *will* fail. This despite years of historical evidence from the Computer Industry that Closed Systems fail. SUN made it hard, it not impossibe for people to take Java and do innovative things with it.

SUN thought that interpreted X-platfrom applications would destroy Microsoft. The only problem was they were trying to replace the Microsoft Empire with another Empire just as Evil.

SUN would not license Java to companies (such as TowerJ) that wanted to build ahead-of-time optimizing compilers (this would make Java just as fast as C.. give or take a few implementation specific percentage points).

They insisted on making Java bytecode interpreted only. The GNU Compiler Collection 3.0 (with GNU Java Compiler) should fix this as it will be the first Open compiler with support for Java. This should remove the last barrier to Java's success! Java will no longer be under the thumb of an ignorant industry giant!

WORA:

It doesn't matter. This has never been Java strength. The problem with JVMs is that they are closed source and impossible to port to other Operating Systems. With an Open Compiler, GJC. This will be no longer an issue because Java will run on every OS where GCC is ported (I think this is basically ever OS in existence).

Personally, I think the creators of Java did an amazing job. The problem is that their implementation was terrible. The Java Language Specification is still an awesome exercise in language design. This work will live on in GJC.

I look forward to the time when developers can work on Free Software in any language they choose regardless politics.

Kevin

11 Jan 2001

Thu, 11 Jan 2001 11:29:23 GMT

Silicon Valley is disgusting. What is wrong with the tech industry? Why is everything about money?

When the industry was first getting started it was about changing the world. Now if you mention you are a Software Engineer people automatically thing IPO, Stock Options, $$, Porshes, and screwing other people over. Gates, Ellison, McNeally.... etc. etc.

What is a guy to do? I am not going to call myself a Software Engineer any more. Software Renegade? Artist? Software Artist... hm.

One thing is for sure. I don't want to categorize myself with unethical snakes that work for the Evil Empire.

........................

11 Jan 2001

Thu, 11 Jan 2001 10:09:21 GMT

Not a very productive day.

Spent some time thinking about the recent Mac announcements from MacWorld (I'm going tomorrow... I LOVE SAN FRANCISCO!!!) and how they apply to Free Software.

The numbers break out to about 7.5% UNIX 7.5% Mac... the rest Windoze. When OSX ships (of BSD with some proprietary crap on top) this should bring a UNIX market of 15% (if you don't consider the other... boring... non Free Unixes). This should be a Good Thing. (don't have to explain why... I am sure you will see :)..

It just bothers me that Apple is SOOO close to getting a clue but they still don't see the light. I can't decide if they decided to base on BSD because they are smart... or too stupid to build their own OS. I think it is a safe bet to assume they are stupid and choose BSD because they couldn't figure out how to write a decent OS themselves.

This leads to why the CHROME (or Crystal... or whatever) stuff is still proprietary. BSD/Darwin allows this. If they based on Linux most of this stuff would also have to be GPL. This 15% I mentioned before might not help us if application developers don't code to Free APIs. Might be a good idea to write an abstraction layer for all of this.

- Why do they charge users $30 for the MacOSX beta... can't you just download it? dumb.

- Why do they only ship MacOSX for their own hardware? Didn't they learn this mistake 10 years ago? stupid.

So again... are they just another stupid/evil company (like Microsoft, SUN or Oracle) or are they actually *smart* but haven't reached the point where they realize that *everything* should be OSS :)

Good Zen quote:

"The best way to control your sheep is to let them roam free."

Kevin

10 Jan 2001

Wed, 10 Jan 2001 13:04:42 GMT

OK.. more SUN rants... Lovely way to patent an obvious technology. This should seriously conflict with XPointer.

10 Jan 2001

Wed, 10 Jan 2001 12:59:54 GMT

Just finished documentation for JavaCore. I have been working on this for ~1 hour every week for a while now (no time). I needed it for Sierra so I cleaned up a lot of things. Works like a charm. :)

It would be nice to implement JPDA support so that Thread stack information, etc, could be provided. It is amazing how clueless SUN is WRT Java. They have such an awesome language but the licensing makes it impossible to use. If they would just pull their heads out of the sand (they still think it is 1994) they would realize they are destroying an awesome technology. (GNU Java Compiler will be a big slap in their face when it ships with GCC 3.0).

Of course I go to look at the source for JPDA (or at least the JavaDoc) and this is only provided under the SUN Community Source License. Yuk. Should actually be called the SUN *DESTROYING* Their Community License.

.. starting to hate this HTML diary interface. Time to write an Advogato posting mechanism for Emacs.

6 Nov 2000

Mon, 6 Nov 2000 12:06:38 GMT

...