Re: Spolsky's requests for Web Applications

You've hit the nail on the head.

Being able to use the browser as a software component --- rather
than turning  it -- anmely the browser -- into an end in itself
is a hole that browsers like Mozilla have fallen into.
Falling into that hole causes one to "call the borwser the
platform" -- and the thing ends up bloating very very quickly
since over time you want "the platform" to do all the things that
other parts of the system are perfectly capable of doing well.

Thinking of the browser as an "embeddable widget" leads to
slimmed down designs that also save a large amount of code --- by
specifically using the "browser" as a rendering engine.

>>>>> "Andrew" == Andrew Mayo <> writes:
    Andrew> To briefly comment on Joel's requirement for
    Andrew> 'contenteditable' (as implemented in IE).
    Andrew> This is a *crucial* feature for building a rich
    Andrew> UI. Absolutely crucial. But the critical thing to
    Andrew> understand about the browser in the context of richer
    Andrew> UIs is that we don't have to just build purist
    Andrew> web-based apps using the browser.
    Andrew> Instead, we embed the browser as a component in a
    Andrew> larger client-side application. This gives us a UI
    Andrew> and rendering engine which - and this is equally
    Andrew> important - can also PRINT things.
    Andrew> I can't begin to tell you how many thousands of lines
    Andrew> of code the browser saves you, as an embeddable
    Andrew> component - I build large commercial systems under
    Andrew> Windows using VB6 as the application language and
    Andrew> embed IE6 as the UI and print engine. Via defined
    Andrew> Javascript functions, the VB6 container interacts
    Andrew> with the UI. This gives you a beautiful n-tier
    Andrew> architecture that is also, potentially, OS
    Andrew> neutral. (because, in theory, I could replace either
    Andrew> VB or the browser without things changing).
    Andrew> To do this properly in the Open Source world I *need*
    Andrew> a browser that supports contenteditable. And vector
    Andrew> graphics (VML is at least reasonably well supported
    Andrew> in IE, for instance).
    Andrew> Unfortunately the Mozilla/Firefox team appear to have
    Andrew> given up trying to implement this feature as it is
    Andrew> 'too hard'. Also progress on integrating SVG into the
    Andrew> standard builds appears to be glacially slow.
    Andrew> Joel, if you're listening. Or anyone who can
    Andrew> help. PLEASE.... Joel, you've got kudos and perhaps
    Andrew> some experiences as an ex-Microsoftee at organising
    Andrew> teams. Can't you help bang some heads together to
    Andrew> make this happen. I am desperate to see this
    Andrew> materialise in a non-Microsoft platform!.

Best Regards,
T. V. Raman:  PhD (Cornell University)
IBM Research: Human Language Technologies
Architect:    Conversational And Multimodal WWW Standards
Phone:        1 (408) 927 2608   T-Line 457-2608
Fax:        1 (408) 927 3012     Cell: 1 650 799 5724
AIM:      TVRaman
Snail:        IBM Almaden Research Center,
              650 Harry Road
              San Jose 95120

Received on Monday, 21 June 2004 12:14:54 UTC