January 2007 Archives

Its the Talk: Some terrible observations on dialogue

| No Comments | No TrackBacks

I've tried writing in my mother tongue. And it's not working. Something is very very wrong in my world. But then I noted that I kind of have the same problems when writing in English too.

Here's the thing: My dialogue sucks.

Fortunately, this is an issue that can be fixed. I've noted that all I need to do is to read what the heck I've written... and revise it. It's just that I really, really, really need to develop some sort of protection mechanism when I look at the dialogue I wrote: It's not necessarily the most beautiful thing in the world.

I guess there's a good chance I'll eventually be able to write good dialogue all the time, but now it seems that I really need revising and rewriting when I write in Finnish. Rusty skills, very rusty...

CMS, your name is pain

| No Comments | No TrackBacks

So, in the unlikely event anyone is wondering what's going on, here's the big deal on why I'm not really writing...

I regrettably get obsessed about presentation. And when I get obsessed about presentation, I turn into a very bad content producer.

So here I am, not particularly producing content... and obsessing over the content management instead.

In October, I wrote on how I planned to turn this site of mine into something that looks like a real site and not some godawful hack. Guess what happened? Forrest turned out to be not yet without its weirdnesses. I tried Lenya, another Apache Cocoon-based thing, but it still leaves me a bit to be desired.

What I really need is templating with XSLT with some mechanisation. Heck, bare Cocoon does it.

Except that Cocoon is a giant monster in itself and really difficult to get started with.

Now, basically, I'd terribly appreciate it if I could get my hands on a version of Cocoon that would actually work if I'd strip the application to bare minimum. It seems that the example app from Cocoon ships with an example web application that has 62 megabytes of stuff. Of which, of course, 11 megabytes are the actual example files. One could make a barebones webapp by copying the 51 megabytes of library stuff... and making sense of which parts of the easy-to-use 2800-line cocoon.xconf configuration file are not really needed.

Ugh. Last time I touched Cocoon, it didn't quite seem this byzantine... It's probably damn powerful if I'd get it to fly, and would do exactly what I want, but right now, my brains are in such a smush that I just can't make sense of all these XML sit-ups.

Someone said someone was working on "Raccoon", which merges Ruby on Rails and Cocoon. Sounds cool and fine. What I'd like to see, however, would be... um... Rococoon. XSLT mechaniser in Ruby based on the otherwise extremely sound guiding principles of Cocoon.

Now, here's my particular list of requirements for a web template engine:

  • Source files should be in any XML variant... roll my own if I need to.
  • Versioning is out of the scope of the engine. As said, source files are in XML. They should be versioned in Subversion.
  • XML transformed with XSLT to produce XHTML out of source files. XHTML transformed further to add a site template to produce the final, static XHTML.
  • Specific requirements: Should be able to produce and preserve comments within XSLT (Creative Commons metadata is added harebrainedly in comments).

Maybe I'll just write a bit of XSLTs and a Rakefile to steer them by.

God I hate some of the code I wrote for my website. And it's really regrettable that Forrest didn't really cut the mustard and I needed to write some really awful horrible hacks to do things that it was supposed to do in the first place. Apache's Java folks write some incredible program designs and then make the implementations extremely heavyweight, puzzling, and add the general veneer of "oh, this is a wonderful app, it just has one small problem..."

That said, I'm trying to cope with my tweak-o-holia and get back to writing even if the website itself appears to be completely messed up.

Where MS Word is Really Wrong

| No Comments | No TrackBacks

Here's some random tired late-night thinking on Microsoft Word. Or, rather, the way Word and word processing programs are used.

The big problem with modern word processing, to me, appears to be that the document styles, document usage, and document versions are being rolled into one. The Document is Everything. The Document Retains. The Document Remembers. The Document Forces.

I have written a doc in OpenOffice.org, but I'm facing the same problem as people who use Word will face. I don't see a solution to this problem because people are so used to Modern, Easier Word Processing.

Here's how things work in Real World: I have written a story. It goes to my beta reader. The beta reader is supposed to use specific character style marking and Insert Note command for their annotations. I even had to make a style guide for this supposedly simple procedure. He's a smart person. He'll figure it out.

Why the style guide? Because a few years before, I Just Sent Them The Damn File... and the user came up with all sorts of funky ways to make their text stand out. Colour the text red, blah blah blah.

What would the printers of yesteryear would have thought if each and every one of the editors would have come up with their own editorial marking styles? They would have been confused, that's what.

Now, I'm a writer of the text. I'm not supposed to be coming up with my own rock-solid annotation method on how to tell apart my own annotations and my beta reader's annotations.

But that's what these new and bold word processing packages force us to do.

Okay, this isn't a big problem. I archive the versions separately. I can read from the svn log that this one here was the copy my friend sent. I can figure out that the red stuff is his stuff, just like ages ago you could tell apart the people who wrote the margin annotations.

The big lesson is this: The new word processing systems don't have a rock-solid support for workflow. They do, however, give you a big honkin' toolkit for defining your own workflow. People who are unaware of how the particular system does their workflow end up in nasty surprises. The whole notion of workflow support comes from the users rather than the system itself.

In a sane world, I could type my text. It'd say "Author's margin notes", "Author's redaction", "Beta Reader X's margin note", etc., wherever such things would actually appear. The Beta Reader would need, without explanation, where to type their own comments to make them appear as Beta Reader X's margin notes.

Now, the Author has to deduce that "this red text here was added by the Reader X, because he was the only one who says he didn't get the copy of the Style Guide."

God, this sounds like some corporate bullshit about changed TPS cover sheets or something. In a sane world, people who write just write. The TPS cover sheets get added entirely elsewhere.

Heck, I'm in so much coffee that I don't know if I already posted about this. But this is how things are...