May 152008

Today Google launched Google doctype, their new open reference for the web.  I’m just starting to look it over, and while I’ll reserve my final judgement until I have the opportunity to spend more time poking at it, I thought I’d share my initial thoughts.

Google is right that there needs to be a solid, reliable, and complete reference to the web.  The Mozilla Developer Center is being built as a complete reference to the open web.  Notice the difference there.  It appears that Google’s aim is to build a complete reference to all web browsers, even the ones that diverge from open standards.

That’s a critical difference.  Mozilla aims to improve the web by encouraging the use of genuine open standards.  The more compatible browsers are with a universal standard, the better the web experience becomes for all users, regardless of which browser they choose.  Our goal isn’t to get everyone to use Firefox — it’s to get everyone using the web to the utmost of its potential.

John Resig has a great analysis of how the tests were done to determine browser compatibility, so I won’t go there.

Reading through the documentation, such as it is at this point, is an interesting experience.  It’s clear that this initial release is something of a rush job, with automatically exported material posted and hoping that people will come along and tweak it up.  For example, take a look at the documentation of the document.getElementsByClassName method in the DOM.  It’s pretty much a stub right now.  The most interesting part there is the test they did to see which browsers offer the method (not to mention the links to John Resig’s blog posts on the subject).

Contrast this to MDC’s documentation for the same property, which includes a description of what the method does, what its result means, as well as a number of examples and a link to the WHATWG specification.

This is a pretty typical example.

Google doctype is an interesting experiment, and I’ll be curious to see what happens with it, despite my confusion over its purpose.

 Posted by at 1:06 PM

  2 Responses to “Initial thoughts on Google doctype”

  1. They could have left an Open Search Description XML, as we have in the (excellent) MDC.
    It would have been more useful

    (by the way, it lacks too in PHP documentation)

  2. Unfortunately, it makes no sense for MDC as it is now to be a documentation repository for the web at large. MDC needs to serve a very different audience as well – extension developers and others who work on Mozilla-based code. This means it must document the Mozilla specific things (e.g. XBL, XPIDL), which is totally pointless and distracting for people looking for documentation on cross-browser things.

    If MDC is intended to document the rest of the world, it needs to have a good filtering capability, such that the user can choose which set of documentation to look at. Ideally, with different URLs (generated from the same set of sources, or something). This needs to exist for different Gecko versions as well (1.8, 1.9, .next at the minimum). MSDN does this very well, with every page noting the minimum required version – though I believe there’s an opportunity to improve by reversing to browsing documentation for the intended version.

    P.S. Not sure if the MoFo installation is better (still can’t see it!), but dekiwiki seems to be worse than MSDN in being totally unusable with scripting disabled. All the menus are DHTML with no fallback.