Mozilla Announces Web Development Learning Initiative 44
bonch writes "Mozilla has announced Webmaker, a web development initiative aimed at teaching the average user the building blocks of the web. Users can join a 'code party' and learn web development with provided authoring tools, and existing developers can volunteer to run their own events. To kick it off, Mozilla is announcing the Summer Code Party starting June 23."
great (Score:4, Interesting)
I love all the new tools out there to help people engage with technology on a deeper level. Going to have to see if we can set up a local event for kids from the nearby school. They have been using Codecademy [codecademy.com] this year in some of the classes, seems like this might be a nice bridge over the summer and maybe something they can use throughout next year.
Re:Oh God... (Score:4, Interesting)
He's probably referring to XHTML, whilst XHTML2 wasn't really ideal, it was at least going down the right path in terms of doing just that. Though it's worth pointing out that Javascript/HTML is itself an example of separation of concerns - client side scripting separated from document structure, though HTML5 does nothing particular to cripple this at least.
Personally I'm not keen on semantics been wedged into the HTML5 spec though, I think that's a concern that could be far better dealt with like stylesheets. The semantic tags are one of my pet hates in HTML5, because they're based on a study that was out of date shortly after it was produced and it was produced in something like 2006. The set of tags is so small and arbitrary that it's meaningless, it would've made far more sense to apply semantic information to classes and ids as that way you could store them in a separate file, which gives you the additional benefits that they can also be applied to existing sites that are no longer maintained in the same way you can apply custom stylesheets. You could have separate people handle the semantics, which would be handy on massive projects, and you'd be able to update the available semantic options without having to update the main HTML spec or have a "living" non-spec as WHATWG have proposed. ARIA roles already basically duplicate the semantic tag functionality but do so in a far more flexible manner, so it's an example as to where, IMO, we should have gone, and should be going.
There are other issues with semantic tags - older browsers just don't even recognise them and fail to render those blocks, you can fix this with Javascript hacks, assuming everyone has Javascript enabled and working of course (they don't). They're hence not even properly backwards compatible, so if you want to use them you've realistically got to accept that someone is browser your site with an HTML5 compatible or higher browser. The other argument for the tags is to get rid of div soup, and sure it's a pain, but honestly? the benefit of divs is that they're generic, and by keeping your tags generic and applying additional information (i.e. styling) using the ids/classes, you're maintaining flexibility and keeping things both futureproof, and backwards compatible at the same time - new stuff can be added whilst old browsers can safely ignore it. Even now the semantic meanings of the semantic tags have become meaningless anyway as there's so much disagreement for example as to what should be an article. There's any number of questions that go something like this "There isn't a semantic tag to mark up this part of my web page, what do I do?" to which the response is "Use divs anyway", or "Use this semantic tag, which isn't exactly what the element is but is the closest element" - the problem with the latter which many people accept and use is that it dilutes the semantic meaning of that tag, an article is already not actually an article half the time, it may be a comment. You could argue a comment is an article, and that's great until you go to the next HTML5 page where someone has instead arbitrarily decided a comment is actually an aside.
There's some nice stuff in the spec, canvas being the obvious one, and web sockets if they ever get implemented properly in every browser, but in general I agree with the GP and think it's pretty awful. CSS3 in itself is pretty nice though, and doesn't have many issues with it beyond it's current level of support.
Really, as I say, CSS3 is pretty decent, but HTML5 is indeed a terribly thought out spec, the whole process has been rather amateurish and felt more like they were building a script kiddies wish list of shiny stuff rather than a spec to take the web forward. Many things such as accessibility and security seem to have been worryingly neglected in a number of areas. It doesn't help when you see grand commander Hixie engaging in flamewars with people and basically ending up with comments that amount to "I don't care what you think, I'm doing it my way whether right or not" - h