Scribd's iPaper and the fragile web

February 22, 2008 ⋅ 3 Comments »

I’ve been taking a break from my RSS reader for the last couple of weeks, so I didn’t hear about Scribd’s iPaper until yesterday. If you also need to be filled in: Scribd is a Y Combinator startup who writes software “that makes it easy to share documents online.” They want to be the YouTube of documents. iPaper is their new Flash-based platform for document sharing:

iPaper is a document format built for the Internet. Like a YouTube video, iPaper documents are Flash widgets which you embed in your existing web pages. PDF, Word, PowerPoint, and many other document formats can all be displayed on the web using iPaper.

iPaper is designed to be fast, light, and simple. Because it’s integrated into your site, iPaper offers a fluid browsing experience, keeping visitors on your site. It has a small footprint, doesn’t require the installation of additional software, and it’s not loaded with superfluous features.

Now I would be the last guy to defend PDF. I’ve invented entirely new swear words just for Acrobat. And I’ve written before about how downloading documents is a seam in the web experience. iPaper seems like it could be an improvement — but it remains to be seen.

But what worries me a bit about iPaper is that if it becomes popular, it’s another step towards a centralized web. One of the greatest things about the internet is that it’s mostly decentralized — a single point of failure can only have a limited impact. If a single email provider like Yahoo! goes down, then only Yahoo! users are affected. The same with my web site: if CNN goes down, it doesn’t have an impact on me. But if your site relies on YouTube and Scribd to serve your content, then you are screwed if one of those sites goes down. (Or goes bankrupt, or gets shut down by the feds, or …)

As we move towards the vision of “utility computing” — where CPU cycles and software are delivered like gas and electricity — centralization is becoming more and more of a problem. The Amazon S3 outage last week brought down quite a few sites. Think about all the sites that rely on Amazon S3 and EC2, Google Maps, YouTube, and maybe soon Scribd.

The web is becoming fragile.


3 Comments:

  1. e - February 22, 2008:

    What the is a "document"? The web was created for sharing academic papers. "Document" content should just be straight HTML 'cause that's what HTML is designed to do.

    If they're just trying to do an online rendition of the printed copy, then they're just reinventing PDF. Their tool doesn't appear to provide any extra features, such as annotation, a mechanism for sharing comments, or sharing a single "documents" across multiple sites. When I viewed their demo document that's supposed to contain advertising, it didn't. It sounds like they can't even get their own crippled featureset right.

    Craptastic!

    As to your concerns about centralization on the web, I worry less about utility computing providers (S3, paypal, etc), and more about the centralization of services: google search, gmail, google documents, google calendar, etc. It's only a matter of time until Google (or Microsoft, or Yahoo!, or Amazon) start violating user's privacy. The impact in China is chilling: Yahoo! has cooperated with a repressive regime and people are in jail (or dead) because of it. If utility computing services go down, then there's a denial of service for their customers, but when privacy is at stake, their can be much more serious consequences.

  2. Patrick - February 22, 2008:

    e:

    I agree that it should be HTML, most of the time. But in reality, there are often many decent reasons why it's not.

    I don't think there's much point in saying what the web was created for, because that was 17 years ago, and the web has long since been twisted for purposes other than those it was designed for. And in my experience, most academic papers are distributed in PDF, not HTML. The reality is that many people have documents that they want to share that are in PDF, .doc, .ppt, etc. They're not in HTML because they were not originally written for the web. For example, from my about page I link to a few of my project reports. I don't want to bother converting them to HTML, but it would be nice for people to be able to easily view them.

    You're right that they are just reinventing PDF. That's the point though. But it also does more -- PDFs, Word docs, Powerpoint slides, etc. are all viewable through the same interface. From that perspective, I think it's a win.

    I do agree that what they have right now seems pretty half baked. I tried to follow their QuickSwitch instructions to turn one of my PDFs into iPaper, and the sample code that they gave me was missing a really important piece, which I only discovered by looking through the full reference. Once I finally got it working, the page was displaying a blank document. Only by clicking though to the Scribd version could I actually read the document, and even then, the font was so crappy that it was basically unreadable.

    So, for the concept, I give them a B+, for the implementation, a C.

  3. Grant Robertson - October 10, 2009:

    I agree. I just learned abut Scribd's iPaper and went searching for info about it and found your post. The first thing I thought when I saw their site is, "So they want me to post all my documents on their site and become utterly dependent on them from now into the indefinite future? Never gonna happen."

    .PDF may have it's problems but I certainly don't intend to get around those problems by handing my life over to some for-profit, possibly fly-by-night company.

    What we need is an open standard that anyone can implement in their own software. Oh, wait, .PDF is already open. So there is absolutely no need for Scribd or iPaper to even exist. What's next, a service where I send them all my clothes and can't wear anything unless the service sends some to me in the mail? Because we all know how hard it is to find clothes in our closets. ...