Sunday, October 31, 2004

Delays in Email Blog Posting

I note that the "EJB 3.0" entry that I posted on Thursday didn't actually appear until Saturday. I sent this in via email, and got no indication that there was any kind of bounce or delay on it. If anyone knows anything about how this is supposed to work I'd appreciate comments.

The value of email blogging is great, because it lowers the bar and allows me to pay more attention to the expression of ideas and less to the posting mechanism, which has held me back in the past from posting the more arduous "articles" on MindView.net. So I'm hoping that the hiccups on blogger.com for email posting are just that, and not an indication of bigger problems.

The problems with podcasting

This article summarizes the issues I was talking about with Mike Levin.

The general problem is not really any different than blogs: they may not have content or form that interests you. The brilliance of blogs is that they are how the "custom newspaper" is formed. I could never quite believe the prediction that you'd go to some big news service who would create your custom paper for you, but I didn't realize that personally choosing blogs based on their writers was the answer.

But the addition problem with "audio weblogs" is that they are linear -- there's no way to skim them. And I don't think that metadata is the answer. I think the answer is the same as with blogging: the creator of the presentation needs to work hard to keep your interest and keep up the forward momentum, so that you not only keep listening, but you keep coming back.

I also think that there will be a variety of formats and different people will be drawn to different formats. Some will like more of the background chatter, and others (like me) want something to listen to during long drives, and for us the "chatty" aspect isn't interesting.

Friday, October 29, 2004

Amazing: EJB 3.0

Today I saw a presentation by Michael Keith that blew my socks off. It
was about the history of EJBs and the new (upcoming) EJB 3.0.

My experience with EJBs has been dismal. Every time I've tried to wade
in and understand them, I've gotten mired down with cruft, and have
lost interest. Some people that I trust had even started saying that in
a few years, EJBs would be legacy technology. On top of that, the way
people have actually been building things seems to have been diverging
further and further from EJBs (e.g.: Spring and Hibernate).

So my expectation was that EJB 3.0 would just be more incomprehensible
cruft added on the already existing incomprehensible cruft. I went to
this talk primarily so that I would hear the history and be armed to
damn EJBs once and for all.

All I can say is: wow. I never could have imagined that something as
big-corporation-driven as EJBs could do such an incredible reversal.
EJB 3.0 is so amazingly different than previous versions of EJBs that I
could look at example code and understand it immediately. I could even
imagine putting an example of this in Thinking in Java (alas, it won't
be out for some time yet, far past the book deadline, and even then
it's not clear when there will be implementations).

The original EJB spec was developed in a vacuum, based on the fantasies
of a group of people about how these things would be developed. The
reversal has come because someone (who Michael mentioned was actually
on the original team) said: let's base the new spec on what people are
actually doing, for example in Spring and Hibernate. And so it goes
from being incredibly obtuse to incredibly simple, from demanding that
you jump through hoops to jumping through the necessary hoops so that
you don't have to. It makes use of annotations and POJOs and relational
databases and all kinds of things that you want to do anyway. And the
best part is that it looks like something I want to play with, rather
than run screaming from as I have with previous versions of the EJB
spec. I actually want to use these things now.

I'm amazed. Kudos to everyone who's working on this.

Wednesday, October 27, 2004

Colorado Software Summit, Wednesday

Noel Bergman explained JNDI and LDAP. He's clearly one of the techies
who is fascinated by the technology in itself, because he showed things
that you could do with the technology. I began pestering him for a
motivation. Not that I think it's useless -- it seems to be a very
clever and well-thought-out solution. But when you look at a design
pattern, which this smacks of, the first thing you find out is the
intent and motivation, so you know where you want to apply this pattern
and what kind of problem it solves.

JNDI appears to be a central repository of information structured in a
hierarchy. As far as I can tell, it can be any kind of information
including serialized objects, but there are formalisms for more common
types of information. Apparently LDAP servers are typically designed
for reading more than writing, so they aren't intended to be the
equivalent of databases.

JNDI/LDAP seems to cut across languages and platforms. In addition,
LDAP servers apparently scale and distribute transparently. All
attractive features. Noel said he was changing a configuration system
that used XML so that it used JNDI/LDAP instead. This sounded like it
could be a good idea, but I couldn't get a clear answer of what his
motivation was to make the change.

So it sounds like a very interesting and useful technology, but I'd
like to know more about the design motivations for using it.

Gary Ashley talked about various open source issues and ways to use it.
To help make his point, he had put Linux on his laptop, and had not
been able to get the mouse working. I love the whole idea of Linux and
open source, but there always seems to be a cost, for example herculean
efforts to get some aspect of the software working. Or difficult and
confusing configuration in general, which seems to be a legacy of Unix,
where you are encouraged to create little languages for everything. Not
everyone knows how to create easy-to-use little languages.

I have to say, Linux has the best screensavers, and it seems to have
continuing evolution in screensaver technology. Windows screensaver
development stalled long ago as far as I can tell.

Jasper Report Builder is apparently an open source way of building
reports. These can for example be output in PDF. There is a plugin for
Eclipse so that you can do drag and drop form building. Might come in
handy.

Tuesday, October 26, 2004

NetBeans: VB for Java?

I was talking to someone at the Colorado Software Summit
during the welcome reception Sunday night and I mentioned
how Java didn't seem to be generating too many desktop applications
considering how long Java has been around. It seems like the
server is the place for Java.

He said he had been using Sun's NetBeans and really liked it a lot,
and that creating GUI desktop apps was very simple. He made it
sound like it was on par with VB, or at least that it simplified
creating the UI enough that it was very worthwhile. Sounds like it's
worth a try.

I wonder if it works with SWT components, as well?

Thursday, October 21, 2004

Commenting is a problem

It appears that HTML pre tags don't work, so posting code is not
possible. That's a problem. Could be pretty serious for me, since a
lot of blog communication is going to be about code. This is
definitely a wall -- It won't even pay attention to indentation.

I don't know if I can use PRE tags in the blog entry itself. Here's a
try:


def spewFile(fileName):
for line in file(fileName):
print line

new ArrayList<String>


RSS

Well, FireFox detects an RSS feed when it's pointing to this page, so
it appears they have this automatically available (although I've never
actually had any tool that does the RSS subscription thing so I don't
really know what to expect). Ding. Another feature taken care of. There
must be a downside to this thing. Although Daniel said Google owns it
so I suppose there might not be, or perhaps it's just very tolerable.

Hmm. They must be thinking about podcasting or audioblogging or
whatever it's supposed to be called. That would be nice but I don't
see them providing the bandwidth.

I wonder what happens with URLs. Here's a raw URL:
http://mindview.net/WebLog
Here's one inside HTML tags:
weblog

Because it does the RSS thing, I could probably remove the RSS from my
site (since it took a lot of bandwidth) and instead put it here.
People could subscribe to this RSS, and then when I publish an article
I could simply announce it here. Also, I could just use the feedback
mechanism they have here.

Who would have thought it could be this easy?

Ok, it's clear that email blogs work, based on the posting below, which I successfully sent through email.

I guess as a programmer I always tend to think of the complexity of
things. And of course I've been bugging Bill about this for what seems
like at least a year, and it's just more than he was ready to do, so
it did seem hard. But now I've spent about 20 whole minutes and I have
something working (not like I actually programmed anything, of course,
just did some searching through the help system).

So the temptation, of course, is to just make this my "casual thought
blogging" area (does anyone really want to read my non-programming
thoughts?). Or make it about stuff my audience wants to read, but less
formal, and with allowances for slightly off-programming remarks. I
guess I'll have to give it some thought now that I'm face-to-face with
the real possibility rather than just imagining "wouldn't it be nice
to have email blogging."

It is awfully nice, though, just to be able to write an email and not
go through the big operation required to publish an article. I find
these "lowering-the-bar" factors can often be extremely important. If
you only have to do "a few extra things" then those barriers can often
prevent you from doing something.

And now for something Completely Different

Email Blogs! Hooray!

One thing that bugs me is imagining these ideas for months or years
instead of trying something out. Of course, I didn't actually *know*
that someone else had implemented this feature, and that I could have
had it up and running in no time.

Of course, I've invested time in my "article blogs" now, and I suspect
I'll still keep those, possibly renaming them as articles.

I'm not sure what the limitations of blogger are. What if I get a lot
of hits? Do they have RSS implemented (I think so). It's certainly
tempting to do my "though blogging" here.

Well, let's just see if this works.