Slashdot Mirror

"Software components of the SLC technology will be made available under an Apache 2.0 permissive open source license, except to the extent that releasing code puts privacy and security of student data at risk .. Applications developed by third parties that are interoperable with but separate from the SLC technology will not be subject to the SLC open source license ". link

"Minimum Application Requirements .. Applications must leverage SLC technologies. Full developer documentation can be found at dev.slcedu.org"

"Last week a subset of the SLC dev team headed north from OSCON to Seattle to host an SLC Camp for about 100 people at the Bill & Melinda Gates Foundation. Our goal was to give members of the team there a chance to ramp on the SLC technologies and their implications for K-12 education"

"Minimum Application Requirements .. Applications must leverage SLC technologies. Full developer documentation can be found at dev.slcedu.org"

"Last week a subset of the SLC dev team headed north from OSCON to Seattle to host an SLC Camp for about 100 people at the Bill & Melinda Gates Foundation. Our goal was to give members of the team there a chance to ramp on the SLC technologies and their implications for K-12 education"

The Course resource looks like an amateur listing everything they could think of, and they still got it wrong. Look at "minimumAvailableCredit" and "maximumAvailableCredit". First, this is just bad data design: either lookup the min/max on the data tables, or if these are proscriptive then there's no way to deal with changes in regulations over time. Second, academic credits vary by many factors (like classroom hours, enrollment types like auditing, etc...) and it'll be meaningless to say the minimum is zero for every course. Third, "academic" credits are not the only type of credits that schools deal with (think lab credits, on-the-job credits, etc...).

This looks like they're trying to build a Cathedral

The below is a rant. You've been warned.

The SLC developer "documentation" was written by bozos who have absolutely no perspective outside of their enterprise clusterfuck swamp. Here's a representative example:

resource - Under the industry standard representational state transfer (REST) software architecture, this is any meaningful concept around which a user interaction can occur.

So, yeah, I get it, a resource may be, um, an argument. Yeah, a verbal argument. I mean come on, try and argue that it's not a "meaningful concept" around which "user interaction" can occur. I mean I'm a user and I can have verbal arguments, duh. Another one:

standard field - A field that is a part of a resource representation, as determined by the schema of the resource.

Dude, a standard field is a field that's defined in the schema of the resource. That's it. Stop with the wordleaks.

The documentation is from someone who can't say what they fucking mean, someone who should have had their fingers slapped with a wooden ruler in their high school writing classes until they fucking got the message. I don't care that they are enterprise geeks who have to deal with various abominations and progress meetings day in, day out. Learn how to write or shut the fuck up.

Sorry, it's this kind of bullshit contentless drivel that drives me nuts, that equally drove Feynman nuts BTW, and for a good reason. RJF hated elaborate abstract frameworks built up around trivial ideas, used for nothing else but aggrandizing the trivial ideas. It's mental masturbation, it's done by people who don't realize (or pretend so) that there are clever folk out there who see that the king is naked, that all those abstractions are built around a single piece of poo in the loo.

Say it like it is. Use common language where such works. Don't wrap things up in abstractions for the sake of abstractions. Sure, I do understand that an API is an abstraction, but you don't have to use a yet another layer of abstraction when describing stuff for crying out loud! And don't fucking make a concept-explaining document something that's split up in a thousand html pages with a couple paragraphs on each! If I'm new to that stuff, I'll want to print it out, spread it out, and work with it. How the fuck do you work with a thousand html files? Do they think they are so fucking important that anyone who wants to touch their heavenly documentation is supposed to write fucking scripts just to collate their driver into a useful form? The only thing missing in their docs is ads. It's make it just as useless as, say, eHow.

It seems like the projects aren't particularly complex, but the barrier to entry is high because documentation sucks and unless you have first hand knowledge with enterprise mental masturbation, you'll spend tons of time figuring out the trivialities that could be spelled out in a 5 page pdf (vs. their idiotic bazillion page HTML thing only available in pieces that pretty much only lack ads to make a complete serving of typical internet barf).

Never mind that their dev website is a typical contentless bullshit "socially driven" page where you can't figure what the fuck the whole thing is about. I mean, they have a freaking twitter feed there. Who the heck needs a twitter feed and pics from, apparently, Times Square, on a dev page is beyond me, but hey, when you lack real content you're free to put up junk space fill, of course.