CSSquirrel

I almost died today.

I wish I could say it was due to running into a burning house to rescue a baby or narrowly avoiding tumbling over a cliff in a high speed chase to rescue a sexy woman from a horrible fate.

Instead, I was eating pizza.

Read the rest of this entry »

This isn’t precisely fresh news, but in August, Justin James cornered Ian Hickson in a HTML5 Q&A at Tech Republic.

Jeff Croft noticed something in the article that stoked his ire (warning: naughty language).

What about HTML5 could be provocative?

It’s Ian’s timetable for the specification, which started in 2003 and he hopes to have finished by 2022.

No, that’s not a typo. For those of you with mathematics disorders, that’s nineteen years total, with the whole shebang completed fourteen years from now.

I’ll be forty-five.

I’m sorry, but that’s bananas! I don’t care that the last ten years of that schedule is debug/feedback time. I find it inconceivable to think that something else wouldn’t have come along by then, rendering HTML5 irrelevant, as today’s comic posits.

Fourteen years ago the best way to view the web was Mosaic. Remember that?

I do. From high school. I was also playing video games on the Super Nintendo and Lucasarts was producing games that didn’t just involve Indiana Jones or Jar Jar Binks.

A Star Wars prequel trilogy still seemed like a good idea back then.

The fact is that with the pace of software and hardware innovation it seems absurd to assume anything about what the web will be like in another fourteen years. I hope that HTML is a memory by the time 2022 rolls around, replaced by something exciting and unexpected that turns the Internet upside down in the same way the WWW did when it was first conceived. Despite all the foot dragging that software giants have caused in the implementation of standards in the past several years, we can’t allow ourselves to grow complacent and assume that we’ll be using the same playbook after enough time has passed for America to finally get off its lazy butt and return to the moon.

Like Croft says in his post, specs aren’t a bad thing. However, I agree with him that it’s definitely absurd to assume that HTML5 will be relevant by the time they hit the end of their projected road.

I’m excited for HTML5, but by the time they put the stamp ‘done’ on it, I predict it’ll be an artifact of Internet history. But then, that seems to be how the W3C works, rather than creating standards they’re just putting the seal of approval on what’s already happened.

That said, some wit has made a convenient countdown clock for those of you that want to keep dibs on the spec’s progress.

Yesterday I launched Robot Rum, an automatically-generated webcomic/humor experiment that involves robot pirates. At the moment it’s more like Mad Libs (holy crap those still exist?) going through a blender with some pretty pictures attached, but in theory at some point Pete-O-Tron (the anthropomorphic PHP script that runs the site) should be fine-tuned to the point where it starts making sense and sounding funny.

Regardless of whether Pete learns humor, the site was fun to build for a number of reasons, one being that it was my own project and not a client’s. So if it was going to be insane, at least it was my brand of insanity for a change. Another is it let me experiment with a lot of goofy ideas or concepts that I’ve been playing around with in my head for quite a while.

Since I can milk talking about the design/development choices that were involved with Robot Rum for blog content, I’m going to do a mini-series on the topic. First up:

Search Engine Friendly Comic

For some reason I keep coming back to working on webcomics. Before Robot Rum and CSSquirrel there was Nervillsaga, a now archived and offline comic following fantasy adventure stereotypes. Like CSSquirrel it had update frequency issues. It also was, as is typical for webcomics, a single image for all the frames, characters, text, and so forth.

For the most part, there’s nothing wrong with that, but one thing that I’ve recently thought is that it also meant that each day’s comic provided no real new content to search engines. It was just an image with an alt attribute of ‘comic’. So after about 600 strips, the search engines still didn’t have any real feel for what was going on other than the little blog-like text I’d follow each strip up with (this was before I knew what a blog was, and apparently still felt like harassing people with details of my life). Now, I could have inserted all the text of the comic into the alt attribute, but I always felt like stuffing attributes with a couple paragraphs sort of violates their purpose, and once again there’s no semantic context for any of the text. To the search engine it’s just one long string.

About a year ago I started experimenting with the idea of assembling a comic by layering transparent PNGs and floating the speech bubbles as actual text. At the time I was goofing around with sIFR as the possible candidate for making the text look appropriate (almost universally webcomics use fancy, cartoony fonts appropriate to the genre they’re in), but there were a number of bugs that cropped out and I became distracted by something shiny.

When I came back to the idea of doing this, with Robot Rum, I decided to jump into the deep end. The comic acts like a stage, with each ‘actor’ absolutely positioned in his appropriate area as PNGs, complete with a little expanding speech bubble that holds their text. When the styles are off, you’ll see that the markup reads like a small script, with the day’s title at the top, the date, then the dialog (in order) complete with who is speaking. In future revisions of the site there’ll be a description field that paints the setting, and any ‘props’ on the stage (such as the crow’s nest) will be better fit into the script (rather than it’s current existence as an arbitrary ‘crow’s nest’ piece of text below the dialog.

What this means is that with each new comic, regardless of whether there’s user feedback on that day, search engines will have more text in a slightly more useful semantic fashion to associate the site with. And as Pete (hopefully) streamlines into a more humorous author, the jokes and gags of the strip are readable in order.

I suppose as a side effect this means the comic will be more accessible to a screen reader, which could actually read the text of the dialog. This wasn’t deliberate, though, and there’s plenty of other aspects to the site which would need to be fixed prior to me making any claims about accessibility.

I doubt that Robot Rum is the friendliest such site for a search engine, as I don’t have any expertise in optimizing other than adhering to standards as much as possible. I haven’t really played around with keyword, but at least its content should be more concrete for robots as well as humans, and if Pete had feelings and thoughts I’m sure he’d want that.

Yarr!

September 19th is one of my favorite days of the year for a number of reasons. First, it’s right about when the seasons begin to change here in Bellingham, and I’m quite fond of autumn. Secondly, it’s International Talk Like a Pirate Day, and any excuse to act like a pirate in public is a good thing.

It also happens to be the birthday of a certain digestively-challenged friend of mine, but I don’t want her ego to get any larger by putting that on the list of reasons.

This year, it’s also the launch of Robot Rum, a webcomic/experiment that details the misadventures of a crew of malfunctioning robot pirates. The experiemental part is that the comic is ‘written’ (read that: randomly assembled) by a PHP-powered ‘robot’ called Pete-O-Tron that will slowly learn what it means to be funny through user-feedback, hopefully transitioning from awkwardly arbitrary to consistently chuckle-worthy.

As the first comic shows, he’s got a ways to go. An initial challenge will be finding a way to string the various pseudo-random text bits together in a way that generates gags or humorous non-sequiturs. At present Pete’s loaded with some Moby Dick, Treasure Island, Lolbot, and a very tiny thesaurus. As feedback is generated by visitors, he’ll be fine-tuned by yours truly to track a sort of ‘DNA’ of what is funny and what isn’t, ideally getting more and more spot on in simulating humor as his database of responses grows.

Sounds really cool when I say it like that.

It’s also a great way to generate a daily product when you’re as lazy as I am.

So get out there and capture a Spanish galleon, load up with doubloons, bury some treasure, drink some rum (highly optional), and check out Robot Rum. If you’re feeling really nice, use the feedback feature for poor Pete. He’s not too bright, and needs the guidance.

(And yes, I’m aware the site’s a mess in IE6. Apparently 1996 browsers and transparent PNGs don’t talk well. I’m currently deciding how mean to be to people that probably still think grunge is alive.)

On Friday the 19th of September we’ll have reached a holiday I hold in more esteem than any Hallmark-sponsored event like Mother’s and Father’s Days (I love my parents, but I hate buying neckties). I speak, of course, of International Talk Like a Pirate Day. Yes, it’s got the dubious honor of being the only known holiday to have been started by a sports injury, but it’s still a great bit of fun, with a lot of “Arrs!” and “Ye matey’s!” and nautical talk going around.

I also suspect it’s a good day for grog, if you’re into that sort of thing.

It’s also a great deal to oficially launch a project I’ve been working on for quite some time now: Robot Rum.

No, it’s not a site devoted to robot-brewed alcohol. It’s an experimental webcomic (which happens to be about malfunctioning robot pirates) that is generated daily by a ‘robot’ (actually, a small web application) named Pete-O-Tron. Being non-sentient, and having approximately five-hundred lines of code, Ol’ Petey isn’t capable of actually drawing or forming its own dialogue. However, it is perfectly capable of smashing together vector art I’ve created with ad-libbed text from Internet memes, 19th and early 20th century nautical adventure novels, and various other sources.

Initially, I expect the comics produced by Pete-O-Tron to be vaguely horrible, with any humor an accidental product. However, I’m going to be incorporating various feedback systems so that viewers can rate a comic in different ways. Pete will store this info and use it to create a sort of humor ‘DNA’ for each comic, and ideally through enough feedback I’ll be able to modify Pete-O-Tron to learn what things are funny and which aren’t.

Gee, it sounds really ambitious and pretentious when I say it like that.

Here’s the short version: I’m lazy. It’s a daily webcomic that I don’t have to update because a robot is doing it for me.

I’m not halting any other projects I’m on (like CSSquirrel), but this one took a bit of startup time, hence the infrequent updates over here.

Arrr!