Scribd's HTML5-Based Document Viewer Is Now Portable, Retroactively Upgrades 20M+ Embeds

Last May, document sharing hub Scribd launched one of its biggest features yet: a document reader based entirely in HTML5. Up until then Scribd had presented its documents using a Flash-based viewer called Flash Paper, which worked well but had a few shortcomings — namely that it didn’t work on mobile devices like the iPhone. Scribd’s solution was to use optical character recognition on a document to effectively recreate its fonts in a digital format, then to use HTML5 to style and lay out the document’s content. And it really works.

Today, Scribd is taking the next step on its HTML5 rollout: it’s swapping all of its embedded document widgets that are using Flash over to a new embeddable version of the HTML5 viewer (and new document embeds will be in HTML5). This is important, because Scribd already has 20 million embeds scattered across the web that will be automatically converted to the new viewer (CTO Jared Friedman says this shouldn’t break anything).

The reader will allow document embeds to be read on mobile devices that can’t use Flash, like the iPad and iPhone, and it should generally offer a more native feel in the browser than Flash does.

The feature has been long-expected — in some ways Scribd was actually beaten to the punch by fellow Y Combinator alum Crocodoc, which launched its own embeddable HTML5 document viewer last month. Of course, Scribd is still a much bigger site, and the services have different use-cases (Crocodoc is focused on making it easy to share, edit, and annotate documents from within the browser, whereas Scribd is a huge repository of documents with more sharing features).

One other thing to note: Scribd actually has a patent pending called “INTEGRATED DOCUMENT VIEWER WITH AUTOMATIC SHARING OF READING-RELATED
ACTIVITIES ACROSS EXTERNAL SOCIAL NETWORKS” that includes claims to inventing the presentation of documents using HTML5 (it discusses the generation of font files from documents and the use of the CSS 3 @font-face tag to present them). At this point the patent hasn’t been granted yet but if it is, companies may have to work out deals with Scribd to offer similar functionality down the line.

I’d embed a sample here, but WordPress still has to update their Scribd shortcode. In the mean time, all documents on Scribd now offer the HTML5 embed by default.