top | item 1327459

Scribd in HTML

315 points| ZeroGravitas | 16 years ago |scribd.com | reply

104 comments

order
[+] pak|16 years ago|reply
These folks just went from my list of websites I dread visiting to "damn, that is some sweet technology".
[+] brandnewlow|16 years ago|reply
And they probably wouldn't have been able to pull it off at this scale had they not first started a site you dreaded.
[+] fauigerzigerk|16 years ago|reply
It may be "sweet" technology. I just haven't found out yet what the purpose of this or the previous technology is.
[+] ZeroGravitas|16 years ago|reply
For a more standard document example, with complex math, charts etc. see:

http://www.scribd.com/documents/5/Paper-5

[+] nkassis|16 years ago|reply
Wouah, that one was really awesome. And in full screen mode, it's as good as reading a real pdf. Awesome work scribd. Performance wise, Firefox seems to have a little bit of a hard time.
[+] hga|16 years ago|reply
Excellent (on my own build of 64 bit Firefox on Linux). I will now start using Scribd for more than just downloading stuff on it.

One note about using NoScript: I can toggle on and off the annoying Facebook popup by hitting the S button it points to, but I must enable a Facebook site to banish it with its X in the corner widget.

[+] dagw|16 years ago|reply
Is anybody else having problems reading that in full screen? Under Firefox 3.6 and Windows 7 it consistently locks up my browser after scrolling 4 or 5 pages.
[+] warfangle|16 years ago|reply
Awesome.

One note: is there a way to pin the bottom toolbar as hidden? It's kind of annoying to read a document with it popping in and out every 30 seconds...

[+] mikelikespie|16 years ago|reply
We're probably going to do something to improve the toolbar hiding interaction. Thanks for the feedback.
[+] maukdaddy|16 years ago|reply
Works really well here on FF3.6. This is REALLY awesome!
[+] papersmith|16 years ago|reply
Another thing is that scrolling is kinda sluggish in Chrome on my 2007 Mac Mini. Google docs PDF reader seems to scroll much more smoothly. It's still a big improvement over the flash version.
[+] newobj|16 years ago|reply
This is a major coup for Scribd. I think it's a service that people begrudgingly and painstakingly used. But now it's just joyful, and kind of sexy. But as someone points out - Scribd is a bridge and if the things that they bridge to ever contain the ability to embed or convert to HTML5, what will they do?
[+] mechanical_fish|16 years ago|reply
Their employees will retire, to live forever on the free pizza and beer which every geek will gratefully buy them in exchange for compelling every single popular proprietary document format to support effort-free export to standard HTML.

Or, you know, they'll invent something else. It's a lot to ask of a business plan that it be good for twenty years.

[+] rev087|16 years ago|reply
Reminds me of the Google Chrome cartoon: http://www.google.com/googlebooks/chrome/ Even the colors, characters and "objects with faces". Not necessarily a bad thing, just curious.
[+] apgwoz|16 years ago|reply
I was thinking the same thing. However, it made me respect them just a little bit more. They could have just written a blog post and said... "We're no longer using Flash," but instead, they took the time to build something that a) illustrated very well their new technology, and 2) actually entertained some of us. Valiant effort.
[+] snowmaker|16 years ago|reply
You called it. That was exactly what inspired this presentation. That Chrome cartoon was awesome - made me an instant Chrome advocate.
[+] flatline|16 years ago|reply
This is really nice work guys. It's slightly disorienting to see HTML5 in real use. I kept looking for a scollbar until I realized it was just a regular web page; the scrollbar is my browser's scrollbar. (Though I think part of it was the positioning of the toolbar at the bottom; it looks like the page is split by the toolbar and it's not really.) The rendering in-browser is beautiful, and much more usable than a pdf viewer on my system.
[+] gamache|16 years ago|reply
I applaud their effort, but when I look at real documents in both formats (e.g. http://www.scribd.com/documents/5/Paper-5 like ZeroGravitas posted here), the HTML5 version looks very noticeably worse than Flash (tested in Chrome 5, Safari 4, FF 3.6). Fonts are rough or missing, kerning is shot to hell, and layout looks like it was performed on a shake table. It resembles the output of a poor PDF viewer.

I wonder if this was the right time to roll out the HTML5 format.

[+] scott_s|16 years ago|reply
I don't know if we're seeing the same thing: http://people.cs.vt.edu/~scschnei/pictures/scribd.png

That's with FF 3.6.3 on a Mac. That looks pretty damned good to me. I like that much better than the Flash version, party because I think it renders better and partly because it's easier to navigate. The only problem I see is that the text could be darker so that it would pop more.

[+] naner|16 years ago|reply
Can you do screenshot comparisons? I am not seeing these problems and I don't know if it has to do with my browser/OS or because I just don't notice these things.
[+] josefresco|16 years ago|reply
Also super laggy when scrolling fast over the content. I never see that kind of scroll lag and it was quite apparent while viewing your provided ZeroGravitas link.

I have to say going from a "viewer" to having all the content on the same page is a huge improvement.

[+] ugh|16 years ago|reply
I was quite impressed by how good it was. There are problems but they are at least not immediately obvious. I’m still not sure, though, whether it’s the right thing to do.
[+] mhartl|16 years ago|reply
The HTML5 fonts look fine for me (FF 3.6.3, OS X 10.5), but the text overflows the right margin rather badly, and it persists even after resizing. I hope they can work out the kinks.
[+] nexneo|16 years ago|reply
Mac/Safari 4 Looks perfect. Better then boxed Flash version
[+] anigbrowl|16 years ago|reply
Fine Job, both technically and graphically. Smooth on Chrome. The only aesthetic change I'd make is that the up/down buttons could jump, or scroll faster - at first I was unsure what they were doing different from my window scroll bar.

I did find one bug...On slide 14 (well actually all of them, but it's most noticeable there) you can indeed highlight and copy the text...but not the last character in a block. If you try to select the last character you'll invert the selection to be from the start of the slide to your highlight point (sometimes this includes the page frame so it looks like you've selected the whole page). I suspect (based on my own bad habits) that it's a boundary error, counting the length of the highlight from 1 when the string length is counted from 0.

Rendering more complex documents isn't as perfect as pdf, eg column-spacing or margins can look a little bit off, but that's a minor cosmetic flaw that I'm sure will be fixed.

[+] jerf|16 years ago|reply
I've noticed Chrome itself has some highlighting issues. Another example: Put some text in on http://www.eeemo.net/ , which is the Zalgo text generator, which works by putting lots of Unicode character decorations on your base text. Chrome demonstrates some very strange highlighting/copy/paste interactions on that page if you try to copy the Zalgo-ed text out.

My guess is that it's not a Scribd bug, and I also suspect there's probably not much they can do to fix it.

[+] camtarn|16 years ago|reply
That is extremely cool. I always found Scribd really frustrating to use - the disconnect between my normal browser use and the embedded Flash reader felt about the same as the disconnect caused by viewing the same file in Acrobat Reader, so I didn't really see the point of it.

If browsers take this onboard as a common HTML5 scenario to be optimised, and it becomes a viable, quick and plugin-free way of reading any document online, I will be very happy :)

Kudos to the Scribd team for the mighty effort this must have taken to implement.

[+] tsuraan|16 years ago|reply
As somebody who dreads scribd links (I won't install flash), this really looks like a good step forward. PDFs are still better viewed in okular, but I could see this being useful for viewing MS formats. Really slick interface, too.
[+] someone_here|16 years ago|reply
Likewise, I won't install okular. Evince is the only game in town.
[+] Janteh|16 years ago|reply
Tried it on my iPhone, only the first three frames show some text (no images) and the rest of the sheets are blank. I believe it is the future though, anyone with an iPad who can try this out?
[+] sgift|16 years ago|reply
I have no iPad, but Opera 10.52 (Windows) has a similar behaviour: The first three frames are there (with images), the rest of the sheets are blank.
[+] maukdaddy|16 years ago|reply
I get a "download now" button on my iPad =(

edit: Works great on iPad. Copy/paste doesn't work for text, but that might be an iPad issue.

[+] tyler|16 years ago|reply
So, it turns out that the version of Mobile Safari on the iPhone doesn't support getBoundingClientRect, which we were relying on. The fix for that should go out soon, at which point it will work on your iPhone.
[+] tomlin|16 years ago|reply
Initially I was expecting the purchase of opinion through the long-played Flash vs. HTML5 gang/mob, "Down with Flash" signs high in the air, binary opinionated nonsense that is so common.

Respect paid to Flash for its use when it was needed.

[+] DrSprout|16 years ago|reply
Was it really needed? Most of the stuff I've seen on Scribd was basically pdfs in an iFrame, except replace iFrame with Flash.
[+] kmb128|16 years ago|reply
How will this change affect API users or those who have embedded Scribd content on their sites?
[+] snowmaker|16 years ago|reply
We are working on a migration path that will switch people over to the HTML versions (unless they don't want it). Are you one of those people? If so ping me directly - jared at scribd.
[+] nitfol|16 years ago|reply
Much better than their Flash interface. But completely broken in Firefox 3.0.x (under Red Hat; can't advance past the first page and there are no pictures) and the fonts are too big under Iceweasel 3.5.9 (under Debian). Until they stop requiring a log-in to download the original PDF so I can view the content in a decent viewer, I'm going to continue cursing every time I accidentally follow a link to scribd.
[+] qhoxie|16 years ago|reply
Would you mind emailing me screenshots of the issues you are seeing? We are trying to document all inconsistencies, and your help is much appreciated.

quin -at- scribd.com

[+] tomjen3|16 years ago|reply
Nice. I have hated scribd for years, because their flash app sucked enormously and I just wanted a PDF.

But this is better than that - now scibd is actually useable.

[+] seunosewa|16 years ago|reply
When viewing flash slides, the scrolling is annoying. I prefer to hit right and get the next page. Can't this be simulated in HTML5?
[+] qhoxie|16 years ago|reply
Sure thing; the view mode switcher is in the middle of the toolbar at the bottom. Select slideshow mode from there and you're set.
[+] aboodman|16 years ago|reply
I wonder how Scribd pulled this off wrt HR. I mean, they must have had a sizable investment in Flash engineering. Did those people leave? Did they just start doing HTML stuff instead?

Back in the day when I was doing webdev, there was a pretty serious schism between webdevs and flash devs, and never the twain met. Maybe this is less true now.

[+] ilike|16 years ago|reply
I dont think Scribd made too much investment in Flash developers. On server side they were using "PDF2SWF" released under GPL by Swftools.org. My understanding is that, initially they were using 'FlashPaper' to display documents in front end.
[+] josh33|16 years ago|reply
I feel like the real accomplishment here is that they are converting documents to websites on the fly - complete with font-faces and image positioning. How do you do that? If that's possible, will we need front-end layout developers in a few years (we'll probably still need animation/transition development, right?)?
[+] latch|16 years ago|reply
I'm totally in that "this is great" camp.

That said, can someone familiar with HTML5 explain, or provide a link, for the seemingly crazy source? Is this a result of the work-in-progress framework that creates these? Or is this really what it takes? Kinda looks like the source out of .doc --> .html conversion.