Article errata

Corrections to a WIRED article titled “Mozilla Brings Webapps to the Desktop, Challenges AIR, Silverlight”:

“Other technologies currently exist for running web apps on the desktop, like Abode’s AIR and Microsoft’s Silverlight. These technologies offer a tight integration with the PC desktop that browser-based solutions can’t yet match. For example, applications using Adobe’s or Microsoft’s frameworks have the ability to operate smoothly without an internet connection, and you can drag and drop things like images and text into them.”

The author is describing AIR, not Silverlight.

(Readable info on SLOOB, Silverlight beta 3′s “out of browser” feature, is available from Tim Heuer… like Prism, SLOOB lets a consumer hack off an existing web asset into a new host and does not provide additional desktop-style APIs. AIR is unique.)

“But while AIR and Silverlight both require proprietary tools to build and run these apps, Mozilla’s Prism add-on uses only the same open-source technologies upon which the majority of the web is already built — HTML, JavaScript and CSS.”

Any Prism page can also be an AIR page — you can use any HTML construction technique for AIR, no “proprietary” tools needed. Same with SWF — you don’t have to use “proprietary” tools to create SWF.

“In the back of everyone’s mind is the idea that the HTML/JavaScript powered desktop apps will soon overtake the proprietary efforts laid out by Adobe and Microsoft.”

In the back of my mind is the thought that propagandists will tell bigger and bigger whoppers until they topple over calamitously.

“Google has also largely solved the problem of offline access using the company’s Gears add-on, which is available for most modern browsers as an free, open-source download.”

Gears tried to fork Apollo’s SQLite APIs, and has not moved out of beta since its arrival in May of 2007. I am not able to readily find web documentation on its current staffing levels.

“The specification for HTML 5 also includes rules for enabling offline data access for webapps.”

As Doug Crockford notes: “HTML 5 is not the next version of the standard. It is a proposal, a work in progress. It has not been through a formal review process. It has not been officially approved by W3C or by any other standards body. Until it is formally adopted, no browser maker should be compelled to implement it. Being subtly incompatible with a working draft is not evidence of bad intent. Indeed, there are some people (such as me, for example), who feel that the whole HTML 5 process is out of control and should be reset, starting over with new rules and better management.”

“When coupled with technologies like Gears and HTML 5, Prism could end up a more appealing, fully open, standards-based alternative for developers wanting to make desktop versions of their apps.”

Prism lets consumers view an existing WWW page in a new browser instance.

Developers can add some authoring hints to these public URLs, to “fork the web” where the same page functions differently in different browser brands.

(AIR does not fork the web, because it uses standard HTML authoring techniques but outside of the normal web addresses that any browser should be able to read.)

Earlier in the article: “The scheme offers a number of advantages, the most significant of which is the ability to sandbox particular web apps. For example if you move Google Docs into its own stand-alone window, an errant script in your main Firefox window could cause your browser to freeze and crash, but your unsaved work in Google Docs wouldn’t be lost.”

This describes running browser instances as separate processes, not sandboxing.

For the article as a whole:

I followed XUL, XULRunner, and WebRunner. Once Adobe started talking about true beyond-the-browser needs with AIR, those efforts were revived and rebranded to defuse the buzz. Yesterday there was a lot of positive attention paid to the New York Times Reader, and today….

“The Open Web” needs to open up.