Shared publicly  -

"So why not build a desktop app for web design around WebKit? I’m not talking about an in-browser AJAX toolkit for dragging elements around and changing fonts, but an actual desktop application built with WebKit as the core to its display. It could have accurate rendering and previews for the way page elements would look, but with some of the WYSIWYG tools desktop design apps have. We wouldn’t just approximate pixels in a flat comp, our CSS would be baked in to the layouts we draw and create on the page. And as Webkit grows, so too could this new app, always taking advantage of the latest and greatest functionality. Just like a browser, it could pull assets from remote servers; and just like a desktop app, it could make use of local processing power and OS-level functionality. This would allow it to effectively combine some of the best of both worlds, with a foot firmly planted in the web.

The advantages would be monumental, allowing a strong creative and explorative process, while seeing how things could react on a live stage. It would fulfill many of the items on my wishlist because these are already part of core browser functionality. We would essentially be designing with live page elements; not a picture of a text field—but a text field you could click into and start typing, and then drag to a different area of the page entirely."
Evan Plaice's profile photoMarsha Frederick's profile photoMichael Stevenson (Sisco)'s profile photoJayden Hudson's profile photo
Looking back, it is amazing that not more has happened in this space. I started using Dreamweaver in 1999 and it still seems to be the best (but not perfect) tool out there.
There was a Mozilla composer (which I think got renamed NVu) that was meant to encompass these ideas but it never really grew after server-side HTML became the norm. What we really need is a webkit/gecko/trident/... renderer inside our development tools so we can use the actual renderers to design the templates that our back-end services fill in.
Mozilla's tool was good, you could switch the editing view between plain text and editable rendered HTML, and I seem to recall you got a two-pane view so you could see your nicely coded HTML and the resulting display as you edited it.
Issue 1: What would the server back-end be

Apache and IIS, I mean, they're good but they're too built to do anything. To make that happen there's a lot of unnucessary complexity baked in so it becomes hard to 'do it right' without opening security holes.

Personally, I'd create an end-to-end solution using Node.js but node is still too bleeding-edge to make it a solid platform.

But how does that integrate with webkit...

Issue 2: webkit needs to be compiled

Like any software that doesn't run on a VM, it'll need a different version compiled for every platform. Plus you'll need a version of the server compiled for every platform too.

That makes it an O(n^2) complex maintenance situation. It is possible but only large organizations have the manpower to manage that.

That could be solved by re-writing the rendering in Javascript to run on Node.js but even if someone managed the monimental task of making that happen...

Issue 3: WYSIWYG editors will always suck

Google did it (Sites), Microsoft did too (Sharepoint) as well as many others. Wanna know what they all have in common? They all equally suck if you need to do anything that breaks the standard functionality of their built-in widgets.

Sure, there are really good drag-and-drop IDEs that are great for building GUIs on the desktop but that's because desktop apps mostly use absolute positioning. The web OTOH uses a relative layout architecture; that's one of the main reasons it's so powerful.

Issue 4: Desktop apps imply stand-alone use

Google and the powers-that-be will never put a serious investment in bringing the web to the desktop because as soon as you disconnect drom the web, you disconnect from their advertising.

Internet market share loss plus advertising revenue loss equals a loss-loss business proposition.

It's not a bad idea but until somebody decides to develop a rendering engine on a platform that already has a good HTTP server platform, there's no point in starting development.

This guy is clearly a designer because he doesn't begin to grasp the issues involved.

It's a simple fact that you will never be able to build a large building without a strong foundation. Us developers devote our lives to building the best foundations possible. Be thankful that you were given the lucury of a powerful design-specific language that's easy enough for a child to understand (CSS) and leave alone us to do out magic on the timeline that makes the most sense.
Great idea, but i feel like it'll be awhile before it ever happens.

The code behind that would be so crazy I shudder to even think of it
Add a comment...