Elf Sternberg

I work with EXT and Prototype at the office, and Dojo at home. I have a lot more time with EXT than either of the other two, so when I’m working on a personal project my first thought is “If this were EXT, I’d do it this way, but how do I do it in Dojo?”

It’s useful to write that knowledge down.

Last night, I ginned up a quick wiki, dropped in a theme I liked, and put the whole thing up. It only has a few entries right now but they should be enough to show what I’m trying to accomplish. If you find this a useful idea, please contribute to it and publicize it:

The Javascript Framework Phrasebook

This entry was automatically cross-posted from Elf's technical journal, ElfSternberg.com

Baling out of politics for a moment (it's gonna be a long few months) but still entertaining myself thinking about how sick and sad this world can be, I'll relate a depressing conversation I had yesterday at Borders Books.

I'd gone to find myself a copy of the Ruby Pocket Reference. I've long liked having paper books for some things; I don't need the tactility of paper for fiction, but for reference it's absolutely critical, especially when I've reached the point of having so many tabs open on the browser that I've lost track of where I put the reference page for CGI and the one for IOStreams.

While I was looking through the shelves to see if there was anything I could learn that I wanted to learn (not generally), I saw a man browsing through the web development section. He was looking at a book on Dreamweaver. Since I'm not a fan of Dreamweaver, I asked him what he was looking for.

"I don't know," he said. "I don't own a copy of Dreamweaver."

We got into a conversation about what he was looking for. He wanted to learn how to put pages up on the web, he said. He didn't know a think about it. I explained my prejudices, and then we discussed the different layers of web development: HTML describes what goes on a page, and CSS describes how it'll look. The limitations of the web server can constrain your site's page-to-page architecture, and an application server like PHP, Rails, or Django (there were Django books everywhere) will allow you to provide pages that react, with things like logins, preferences, per-user views.

He seemed a little overwhelmed with the depth of knowledge needed. And then I pointed out that I wasn't a graphic designer; the visuals I usually just adapted from OSWD or a similar site, or bought outright from one of the many template stores available online.

And then I finally got around to asking him why he wanted to know.

"I thought I would put up a site for different cities that had deals on cars and other things, and would let members know right then that if they went down to some dealer somewhere they might find the right salesguy having a good day, or the right store with a very low price. You know, same day kind of stuff. I heard that there were these sites that had, what's it called, Adwords, down the side and I heard about these guys who put up a site and it made two or three dollars a day, which doesn't seem like much but if I put up a hundred sites like that it would make a lot of money, wouldn't it?"

I sighed and explained that it doesn't work like that. Only a fraction of sites make even a dollar a day. I said that there's a limit to automating the kind of site he was describing, along with policing it for griefers, spammers, and porn hackers (I had to explain griefer to him). What he really needed was a compelling difference from the rest of the market, how to stand out from the noise, and make your site different from everyone else's, something that would generate a lot of interest. "You've got three options: create content, which makes people want to read what you've got to say, create a community where lots of people want to read what each other has to say, or create an application that makes each and every user's lives better or easier in some way."

He seemed very disheartened by the responsibility of running a business. But then he brightened. "Some sites are just collections of links to other stuff. And they make money."

"Yeah, they cheat. And Google's figuring out how to kill them, because they don't actually add to the sum knowledge of the world. They just steal. Someday, Google will cut them out of the search field."

He kinda shrugged and said, "Well, there's still room for people like me, I guess." He pulled out the Dreamweaver book and walked away. "Thanks for the talk."

Sigh.

Well, I finally decided to knuckle down and work on my Rails project a little more. Total development time so far: about two hours. Progress: You can now add genres to the database. You can create stories, and stories have plots, and plots have scenes. You can't create plots or scenes yet, but you can add new stories and assign them genres. The genres drop-down is still just a label; the entire genre annotation subsystem hasn't quite gelled in my head and I think I'm going to have to hack it straight out of reality the hard way, by discovery rather than invention. I haven't even begun to consider the relationship between scenes, plots, and the timeline, but I have some intriguing ideas.

The generic story collective is called "world," but it has no model yet. Eventually you'll be able to add characters and locations to the world instead of to a story so you can keep using those characters and locations over and over again.

Once world is solidified, it'll belong to its own collective, "author." I hope someday to put this thing up on Pendorwright and let people beta-test it, joining as authors. Paying customers will be able to have more than one world, more than five stories, more than 25 scenes and more than 200K characters per story, and will be allowed to add private genres to the genre annotation subsystem. I'll make sure to run diction&style on all new text objects uploaded to the database to make sure that nobody's using it as a generic repository of garbage. Users will always be allowed to download their efforts.

I tried a downloadable "writer's tool," Writer, and I found it bewildering. It's not obvious to me how one "starts a story." I'm hoping that the "New Story" button in my thing will immediately lead you through to the Rails text editor tool intuitively.

I'll also be putting a stripped-down version of the source code up on the site for free, so anyone who wants to can run a local instance. The stripped-down version will support a single author (i.e. it will have no authentication or payment systems) and will lack many of the planned text-rendering features (mostly because they'll be subsystems that rails will call, rather than being part of the rails application itself).

Ruby is officially cool. Even without Rails, I can totally see how this language could become addictive fast. I barely speak the language and am desperately thumbing through the nutshell book at warp speed and I manage to write a better, faster, and more secure version of Narrator, this one obeying the principle of "configuration, not integration!", in less than half an hour.

That's totally cool. The only question remaining is simple: does the lack of an interim on-disk format cause a performance hit compared to the python version?

... will I remember what I'm doing when I'm sober?

The following is a diff entry against the emacs client lj-update 3.3 that supports tagging. Note that the client makes no effort to check the tags, ensure they're consistent with anything else you've ever done, or enforce comma-correctness. Brad's entry makes it clear that tags can be "short phrases," therefore it's up to you to put commas between your "short phrase" and single-word tags.

( Diff follows )