[Axiom-math] RE: [Axiom-developer] axiom, browsers, and crystal

[Page, Bill]

On Tuesday, June 21, 2005 12:17 AM Tim Daly wrote:

> i am trying to avoid 1.2 million lines of javascript with browser
> dependent ifdefs.  we already have that problem with C.

I agree. For desktop use we can easily standardize on one browser.
I recommend Mozilla's FireFox. Except for jsMath which is provided
by another project, I don't expect a browser-based interface to
Axiom to require more than a few hunderd lines of JavaScript.

> instead of trying to work "on top" of the browser, which limits our
> abilities to the current available set, i'm trying to think of it
> as a large library of code we can exploit in new ways.

Because of the DOM api and extensions, it is possible to work with
the browser "as a large library of code" without having to get
inside.

> and the careful choice of a lisp available API would make the
> browser more useful as a front end and be less maintenance in
> the long run.

As pointed out in another email, it is possible that a lisp API
for the Mozilla browser will be implemented by other developers.
If it exists we could in principle use it instead of JavaScript.
I would expect it to be some time before it was a stable as the
current JavaScript API.

> we want to do everything the browser does (because people use
> them) but locally we want to do so much more. restricting ourselves
> to require a network and vanilla browser code severely limits the
> possible creative choices.

For desktop use no network is required.

I would not characterize Mozilla as simply a "vanilla browser". It
has a lot of provisions for plug-ins and extensions.

For access to a Doyen-like server over the web, I think we may
have to be more flexible with the choice of browser.

Regards,
Bill Page.