[whatwg] Browser Bundled Javascript Repository

[Tab Atkins Jr.]

On Mon, Jul 13, 2009 at 3:20 PM, Aryeh Gregor<Simetrical+w3c at gmail.com> wrote:
> On Mon, Jul 13, 2009 at 3:21 PM, Tab Atkins Jr.<jackalmage at gmail.com> wrote:
>> Yes there are.  jQuery doesn't put out very many updates, and I don't
>> think most people alter their jQuery file.
>
> It's only one file, not multiple?

Yes.

> And people don't inadvertently
> change newline style, or minify it using different incompatible
> minifiers,

As jQuery serves a minified version directly, probably not.  I suspect
that most of us just download the file and pop it straight onto the
server.

> or combine it with their other files so that they only
> serve a single JS file?

This is a possibility, but I don't know how often this happens in practice.

> And you're taking into account that certainly
> the majority of sites don't use jQuery at all?

Was that ever in question?  The point is that jQuery (and other
libraries) are used *often enough* that caching across websites can be
useful.  This is the whole reason that the Google-hosted version is
useful.  Automatic caching from the browser just allows smaller-use
libraries (and plugins for those libraries) to benefit without having
to convince a central authority to host them and donate bandwidth.

> And that you'd have to
> remove it when the user cleared their cache?

I'm not certain what this has to do with the proposal.

> Does anyone have statistics on how useful this would be in real life?
> I suspect only marginally.

For starters, anyone know the usage statistics on the Google-hosted versions?

>> That'd only address the things that the various vendors can agree are
>> 'standard' enough to push into a common API.
>
> That would be enough for most people.

The issue is getting everyone to agree on enough.  I use quite a lot
of jQuery's tools, for example.  Even if a lot of it was pushed into
the browsers, I'd still probably have to use jQuery for some of the
functionality that wasn't included.  Of course, at that point jQuery
could potentially be much smaller.

And what if the API ended up resembling Prototype?  I *greatly* prefer
jQuery over all the other frameworks, and would have a much more
difficult time using anything else.

>> It would also mean that
>> legacy browsers break completely, as you're depending on the browser's
>> built-in libraries that don't exist in older versions.
>
> No, because you could have compatibility libraries that you just
> wouldn't load if the browser already supported the features.

Yes, so now I have to feature-test those things before trying to use
them.  Dealing with legacy javascript support is annoying enough right
now - I'd prefer that these things gradually migrate into the core and
the library I use still just wrap around it.

The fact that standardized javascript features tend to end up being
clunky due to their requirement of language independence makes me not
really want this either.  I much prefer a library that presents a
good, clean language to me, and handles the clunky aspects in the
background.  I doubt we'll ever have a standardized AJAX function as
clean and easy to use to use as $.get(), frex.

~TJ