The Goldfish Chronicles

I love this article: http://www.paulgraham.com/javacover.html.

It is primarily aimed at explaining the hacker’s radar, the concept that a hacker can sniff out a bad technology/thing. However, I think it is utterly accurate of Java.

I do like Java as a language, it has a lot of niceness. There are many aspects of Java that I wish other languages had. However, I think Paul’s instinct of the hacker view of Java is spot-on. I am a hacker, I hack Java. The Java hacker community is a smaller cousin of the large corporate community.

Joseph Ottinger and Lance both commented to my linking to Paul Graham’s article. I was replying in the comment bit, but it got long, so here’s the whole reply:

He begins by saying he has no clue about the technical issues of Java, that what he is trying to display is the idea of sniffing out a bad technology.

Also, bear in mind that the article is 2 years old. So a slightly different world. Indeed, before Java ‘hackers’ were as large a group as they are now.

#1 Hyped: I think this bears merit. A technology which is hyped by your manager before your colleagues smells dubious.

#2 Aimed Low: This is the one that irritated me. But probably because I like to think that I can at least compete intellectually with the majority of us programmers out there. If not grammatically.

#3 Ulterior motives: The open-source crowd originally hated Java for its Sun ownership. Sun continue to show that Java exists for the large corporates, not for the tiny developers.

#4 No one loves it: I don’t know. I _prefer_ Java. But I don’t love it, not in the way that perl people I know love perl. A lot of Java developers I know are that way, they don’t love Java, however there’s nothing else out there better. I think most of us are waiting for Java’s successor to finish the work it started. I used to love Rebol, but it was owned by one company and was never freed. I used to love Perl, but it doesn’t seem to support a large enough industry of dedicated programmers.

#5 Forced to use it: This is a bit dated. It’d be ‘forced to use C#’ now. As far asVenture Capital goes, Java seems a bit like Oracle, it makes the VC wet.

#6 Too many cooks: The JCP does seem this way. I imagine that behind the big locked doors of the JCP, the big companies, Sun, HP, BEA, IBM must be in constant wars.

#7 Bureaucratic: I don’t get this one. I’m not sure if he means Sun, the APIs or the effort needed to get ‘Hello World’. But it would fit anyone of those topics.

#8 Pseudo-hip: Well, it is. Java has continued since 2001 to portray itself as a middle ground between MS and open-source.

#9 Large organisations: They rule Java. The JCP is dominated by these, Apache is accused of being dominated by these.

#10 Wrong people like Java: Difficult to accept this one.

#11 Sun are screwed: Here we can all happily go “But we have Daddy IBM!”. Which I don’t think is good. “Free Java!”

#12 DOD Likes it: I think this is some old ADA/Lisp thing coming back. Unsure.

Overall I like the article, and Paul Graham’s articles in general, because they make me think. It accurately portrayed a lot of how Java looks to me [with a few exceptions].

Well, I got myself back on javablogs.com. For a month or more now it’s not been checking my site at all. Get’s kind of depressing

So I finally deleted my entry at javablogs, re-added it, and suddenly it imported 10 or so of the latest entries and splattered them on the site.

It’s nice to feel noisy again. Looks like Patrick Chanezon just did the same thing.

Andy has some interesting things to say about life at Apache. I agree with him somewhat and am dissapointed in some things he’s been told. Such as not to blog about Apache.

Andy is a member of the ASF, so maybe part of his being a member means he has to be an amployee and not discuss his work on his blog. Fortunately I’m not, so I can say what I like.

There is definitely a lot of disagreement between the people on the lists at the ASF, just as there are in all walks of life. An issue the ASF has is a very hands-off management approach to solving disagreements, so much so that often decisions just aren’t made and things are allowed to quietly being anarchic. Take the website for example, it’s a mess. Any company with such a website would garner snide remarks. Apache’s site is huge, it has radically different looks and feels, with one old l&f and two or three new l&f’s.

Another disagreement close to my heart at the moment is in Commons Lang. Discussions on the Commons list between Commons developers are split on whether Lang should continue to grow, or whether new things like math, time, reflection sub-packages should be their own projects. The chief disagreement is over the functor sub-package.

Ultimately, Apache is about personalities. There are definitely some strong ones there. Andy himself has been quite a firebrand. He likes new ideas, and he believes in the Kevin Costnerism of “If you build it, they will come”. This works sometimes, but other times you get slammed. Another big personality is Jon Scott Stevens. A guy with a lot of useful things to say, but with an argumentative streak a mile wide. I think the most respect goes to Costin Manolache, though possibly because I equate him with a friend of mine who muds with the name of Corvin. Costin also argues hard, but I’m often surprised each time to read his fervent rants and realise that they make complete sense.
Yet another is Jason van Zyl, the head of the Maven project. Another who argues aggressively and portrays a strong character onto the lists. That strength has helped Maven retain its focus I think. That might be one of the problems with Commons Lang, it lacks any owners, just occasional maintainers.

There are many more. Many quite quiet coders who are moving things along on their own. Other loud voices who use bandwidth up with their arguments. Dion Gillard’s constant cvs commit messages, living in Australia must help supply time for coding. Or maybe it’s a summer/winter thing, in winter we code less? Though to be honest, I’d have thought I’d code less in summer. Unless it’s tied to moods/morale.

Will be interesting to see what happens to Andy next.

So was everyone aware of this site: http://www.sunsource.net/??

A wonderful article detailing why Windows is still not ready for the desktop market yet.

Got some i18n work coming up at work, so I’ve been a good boy and am setting up a French, German, Dutch [as our primary consumers] and Hungarian [as a non-Latin-1 character set that I can at least pronounce/read bits] Windows versions on one machine.

While ensuring that Hungarian was definitely not Latin-1 [it’s Latin-2], I found this interesting site.

Wiley seem to be pushing a lot of new books over the next year. To list some of the upcoming books:

Apache Tomcat Bible
Jakarta Pitfalls: Time-Saving Solutions for Struts, Ant, JUnit, and Cactus
Mastering AspectJ: Aspect-Oriented Programming in Java
Mastering Resin
Mastering Tomcat Development
MySQL and Java Developer’s Guide

Most notable to me, the MySQL book, by the author of the MySQL JDBC driver, post his joining MySQL AB.

Wiley have 2 or 3 older Open Source Java books too. Generally I’m not a big fan of Wiley books, perhaps because so many of their books seem to look like big fat dumbo books. Hopefully these will be better, their Ant/Cactus book was good.

I try to avoid talking politics on the same forum I talk shop on. But sometimes there’s an important point and you want to make it somewhere:

A quote from BBC:

But US congressman Rob Portman, while ….
He also defended the US’s right to fight for freedom.
“Not just freedom, but democracy. Not just democracy but human rights.”

This is wrong. Democracy is not something to be protected. Governance by a majority preferred entity. Lines like Rob Portman’s above suggest that a nation ruled by a monarchy, in which the people are 51%+ behind that monarchy, is somehow morally wrong. If a nation’s people want a dictator, they should not have democracy forced on them because Rob Portman thinks that it’s the only viable government. If a majority of US citizens became disenfranchised with the US’s brand of democracy, then it would become an immoral form of government for that country.

Steve’s comments got me thinking on a tangent I like to use to hammer home why open source is better than commercial.

It goes like this, and is based on the fact that computer programmers are craftsmen using engineer methodologies to match the demands of the overlords [managers].

Which product do you expect to be better from a craftsman? The lovingly created piece of art that they made for their own home, or the one they made for the wealthy customer who is demanding it be done tomorrow.

The answer is that it depends. Mainly on whether the wealthy customer is supplying resource that the craftsman cannot provide. Therefore a goldsmith’s house is not full of gold. An Oracle DBA is unable to churn out open-source based systems sitting on Oracle. However, if the wealthy customer is not supplying any extra resource, just what the craftsman already has, then I’d much rather have the one the craftsman makes for themselves.

Metaphor meaning: Rather than buy an expensive piece of code that some programmer of unknown quality churned out to a tight deadline, use a piece of open source, and hopefully free-beer, code that a craftsman is putting their reputation on the line for. When did you last see an internal Oracle developer putting their name on the line? When did you see a MySQL developer putting their name on the line? I believe the latter is more common than the former.

There are other considerations though. Product and support.
* Companies write a product that they think a customer will want.
* OSS Developers write a product that they want.
* Good companies provide a support structure to Developers, testing, user-analysis, bug-report handling, user-complaint handling. In a good company, this makes the developer’s job easier and I believe, the product better. Few companies have the resource to be a good company it seems. More likely the support is non-existent. Testing is limited, bug-reporting is left to developers, user-analysis is a bunch of management guesses and user-complaint handling is dealt with by clueless tech-support front-line with printed up notes.
* OSS Developers support directly. This breaks down when user level exceeds developer level. The solution here is that developers stop supporting all users until some of those users can be classed as developers and the system is rebalanced. Commecial companies can upset this by providing their own support, as it can send an OSS project to stagnation or development collapse.

There is one other big difference between an OSS product and a commercial product. OSS doesn’t die. When a company dies, its product vanishes from the market and all is lost. When an OSS project dies, most of the time the site will sit there for a couple of years untended. The project may be restarted by a new individual at any time. Sometimes the bandwidth vanishes, but as there are numerous copies of said code around the world, chances are that it can be renewed there.

So why use OSS? Quality [the coders care]. Stability [it’ll be around forever].

Why not use OSS? Lack of necessary features [not what the user wants]. A tiny domain [few potential users, so developer loss hits the balance hard]