The Goldfish Chronicles

Just a small note that I’ve put simple-jndi v0.4 up online.

It supports XML underneath now and a multiple properties format [ie) Lists in .properties by repeating the key].

I’ve now used it in a codebase as the configuration system, and it was pretty nice, and I’ve also quickly thrown it into use for a simple IDE project loading and dumping from the database, and it successfully fulfilled its number one goal of making it easy to configure DataSources.

Seems we have coding standards for JSP now.

*grumbles at not being allowed to have .inc files for includes*

The net is getting really tired recently. Complaining that slashdot is useless has gone beyond pass頡nd it feels quiet in general. There are a few sites recently that have really perked me up:

Bookpool’s bookmark contest is one such site. I like buying from bookpool, and I even got some of these bookmarks without realising they were special. The Internet helped make this happen.

Billy Bragg provides and authorises mp3’s of his work that was not released through ordinary channels on his personal site and an australian site. This is how mp3s were meant to be released. Stuff I can’t buy in the stores, that is great. I might have to hunt down some albums now [if I can find any over here].

Liverpool FC [yep, I’m a supporter]. http://www.liverpoolfc.tv is a site that for 40 quid a year lets me watch highlights of games, hear commentary of games and watch tv about past years and the youth squad etc. It does pretty well out of it as I tend to forget about it for months at a time [I’m a forgetful supporter] and then slam it to watch the last months worth of action. Mainly I forget because I’m at work while games are being played. The Internet is great for this

Recently I’ve been scratching private itches. Mainly simple-jndi, and a hopefully nice scraping engine that is maturing into a datafeed engine. Quite hooked up to requirements at work, but I have high hopes that there won’t be any problems in open-sourcing the engine.

It’s been fun, one of those things you can come home to at night, move boxes a bit on the diagram, and tweak some code. Just let it naturally evolve itself with new requirements, and not be afraid to take a little while to rethink a design.

Interesting (ish) series of posts on slashdot about prejudice against scripting languages.

It’s quite amazing. I was originally an LPC [Pike] and Perl coder before moving on to Java. I dabble in php and use bash/sed/awk etc a fair bit. Of them all, it is as a Java developer that I feel most persecuted. There are four communities it seems, major ones.

There’s the Microsoft world. As an MS coder, I imagine there is a lot of persecution from less powerful groups, ie) the open source people. Somehow I doubt they are that bothered.

There’s the C community, and C++. There’s Linux, GNU, QT, FreeBSD, and despite all their arguments they exist in the same world.

There’s the scripting world. Perl, python and php. The three p’s. Ruby is here now, Pike is here, Tcl, lots of others.

And then there’s Java. It’s the ugly poster-child that each other community dislikes.

It’s corporate, under Sun, so C/Script people dislike it, and yet it’s multi-platform so MS dislike it.
It’s compiled, so the scripters dislike it.
It’s interpreted/slow to start, so the compiler people dislike it.

You can go up to an MS fan, and get abuse for liking Java, and you can walk over to a C or Perl coder, and get abuse for liking Java.
Most Java developers I know are not wildly enthused with Java as a perfect language, but they’ve also not found a language they prefer. The reason for that might be that it has a lot of compromises of the other communities.

Back to the subject of this entry, I don’t think Java is a compiled language like C or C++. It’s a scripting language, like perl/php/python. Its culture is far closer to the scripting cultures than others, but that may be because i am a scripter in Java. It just happens to compile to an interim format.

And still we wait to learn if LGPL is:

1) Exactly the same as GPL when applied to Java
or
2) Useless when applied to Java, allowing us Java coders to utterly circumvent it.

I think #2. Here’s my reasoning:

First we look at section 5 of the LGPL. This mainly states that as long we don’t make an ‘Execuitable’, then we can link to that work under our own licence. Great. Java rarely has executables, so most of that section is void and unenforcable.

*********
5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a “work that uses the Library”. Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License.

However, linking a “work that uses the Library” with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a “work that uses the library”. The executable is therefore covered by this License. Section 6 states terms for distribution of such executables.

When a “work that uses the Library” uses material from a header file that is part of the Library, the object code for the work may be a derivative work of the Library even though the source code is not. Whether this is true is especially significant if the work can be linked without the Library, or if the work is itself a library. The threshold for this to be true is not precisely defined by law. “
******

5 does however go on to say:

********
If such an object file uses only numerical parameters, data structure layouts and accessors, and small macros and small inline functions (ten lines or less in length), then the use of the object file is unrestricted, regardless of whether it is legally a derivative work. (Executables containing this object code plus portions of the Library will still fall under Section 6.)

Otherwise, if the work is a derivative of the Library, you may distribute the object code for the work under the terms of Section 6. Any executables containing that work also fall under Section 6, whether or not they are linked directly with the Library itself.
*********

Well. This is nonsense to me. WTF is an Object file? never heard of one [the C part of my thinking is currently bound and gagged and listening to barry manilow]. Again, no executables. So really I’m having a hard time understanding what they’re going on about here. Section 5 is pretty much useless.

Sections 1-4 are harmless by the way:

1) You must provide source and can charge people for the source/binary’s transport.
2) Concerned with modifying the LGPL code.
3) You may GPL LGPL code if you wish.
4) Source must be with binary, but it can just be in the same location as binary.

Continuing on:

5) Dodgy one about linking from other projects that doesn’t stick in Java. Parts of it drop through to 6.
6) More linking stuff and relicencing issues. Largely similar to the ASF licence anyway, you must specify somewhere that this product is used by your product. Must include source of this product, use a linking mechanism [errrr Java? :)], provide a written offer [this is the Internet age guys], Verify the user has received. Mainly it’s pretty bullshitty. I can see how it might stop my company selling an LGPL product, but it seems to fit any other open source licence.
7) You may put this work side by side with another work under a different licence. Erm. Bit odd. Side by side?? On my desktop? My website?
You may not go beyond this licence, catch-all statement.
9) You don’t have to accept this licence, but you can’t play ball if you don’t.
10) This licence passes on whenever you redistribute this work.
11) If someone sues you to break this licence, we still own your arse.
12->16) Legal shit.

Now… while I can see that a C/C++ programmer, in their painful world of compile-time linking might have an issue with many parts of the licence, it seems to ignore the late linking that everyone else in the open source world seems to use [java/perl/python/etc].

At the very worst it stops us modifying LGPL code and releasing it in an application. We’d have to modify the LGPL code, release our modification as an LGPL work, then cross-link to that from our BSD licenced work.

Poolman appears to be making a bit of a comeback. This project has taken the poolman source, fixed some bugs and taken it in a bit of a new direction [reported at “>JavaLobby].

Also, the slow-moving db.apache.org is rumoured to be hosting poolman [submitted to Jakarta a fair while back I think, but it got lost in the creation of db.apache.org along with OJB].

[yapoolman link updated. They don’t have an sf homepage yet]

I’ve had a few thoughts over the last few days, but no time to blog. Wireless network playing up yet again.

One of those is that the JFC [Awt/Swing/sound/etc] are a major pain in the arse. As far as I know, implementing them has taken the largest amount of effort for Apple, and BSD are sitting there without a good JVM. The problem being, BSD doesn’t really want JFC, it wants a good stable J2SE JVM for running Tomcat etc.

If JFC was not a part of J2SE, or if there were some subset of J2SE which could be appliable to servers [I’d still expect awt.Point, awt.Dimension but not much else] and make it a lot easier to have Java running on different platforms.

JFC is all well and nice, but it’s definitely the least useful part of J2SE for a server.

Gah. The BBC News site has been updated and feels much more like a CNN style site now. Ick. At least it still feels cleaner than CNN.

Okay, it’s been on slashdot so it’s hardly news, but this note from David Stutz is notable to me in that it is a solid criticism of MS without being a praise of OS.

It always seems that the only sane group amongst the MS marketing, MS engineers and MS fans are the MS engineers. In fact, they seem to have a lot of fun.