Hacker News new | comments | ask | show | jobs | submit login
Ask HN: Review my startup - Scribd (YC S06)
220 points by snowmaker on May 6, 2010 | hide | past | web | favorite | 150 comments
In the past when Scribd has come up on Hacker News, a handful of HN readers made no secret of the fact that they saw the Scribd product as flawed. While the harsh feedback was painful at times, it was important for us to hear.

We are sincerely grateful for the advice and criticism that we got from the very smart people on HN. The conversations we have had with people here were instrumental in our understanding what was bothering people about Scribd, and how to fix it.

Today we're launching the largest product change ever on Scribd, and one that goes directly to the heart of those concerns. We are leaving Flash and moving to all HTML5.

The #1 problem people had with Scribd was the Flash reading experience. Scrolling was problematic, you said. It didn't respect your browser's search, text selection, or keyboard shortcuts. It didn't feel like your PDF reader, either, and it was missing functionality you expected in a document reader. Sometimes it slowed down or crashed your browser.

At a higher level, people challenged why we were using Flash to display documents in the first place. A lot of you suggested we build a Google books style viewer, using AJAX and images.

We ended up taking that suggestion, but we took it a step further. Rather than building yet another reading application based on images, we used the magic new stuff in HTML5 to convert documents into fully native webpages. It took us a while because we built this from scratch, and it was hard to do.

The result is that the reading experience on Scribd is now incredibly light and fast. It's no different, really, from reading a blog, since it's all in plain HTML yet it still preserves the exact complex and rich formatting of the original document. While it's an early beta product, I think it is already way better. But don't take my word for it - try it yourself:


What is Scribd, anyway? There’s been a lot of confusion among HN readers over what Scribd is supposed to do, and I want to clear it up. The point of Scribd is, most emphatically, not to replace your PDF viewer, and that’s why we encourage users to download documents in whatever format they want. If you try to think of Scribd as an Acrobat/Preview/Foxit competitor you’ll never understand it. The purpose of Scribd is two-fold: 1) For authors/publishers: to provide a place where they can publish their book, essay, presentation, magazine, legal document, poem, catalog, scientific paper, etc. to a wide audience 2) For readers: To provide a place where you will occasionally find interesting content generally not available elsewhere and make it easy to share that content with your social graph.

Scribd’s real value is in the unique content – not available elsewhere – that we have added to the web. It is that store of content and the community around it that has made Scribd popular among authors, students, business professionals, musicians, cooks, companies, publishers, and other wide-ranging types of people.

What HN made clear was that while building out that community, we could not ignore the reading experience itself. When you find an interesting document on Scribd, you expect to be able to read it in an easy, natural way. That’s what we’ve tried to build.

Feedback When we've gotten feedback from HN, it has ended up dramatically improving the product. I have every reason to expect the same will happen this time. And I think we've shown that we do pay attention to what gets written here. So, please, tell us what you think in the comments (or email me directly). Kindly keep the discussion to the reading experience itself for now – that’s what we’re really trying to improve.

What else can we do to make Scribd the best reading experience on the web? The whole team at Scribd is anxiously waiting to hear your answer.

For comparison, here's the same pdf as viewed in Googles pdf viewer: http://tinyurl.com/33339j8

My own feedback so far: Clicking download, should allow me to download the document. It should not require me to create a dummy account, jump through 'invite friends' 'look at these other documents' hoops.

Also I really like the way Google has a scrollable list of thumbnailed pages. Makes navigation very easy. Also search, and ability to change size of pages with a simple click to allow different reading modes. Plus with Google, when you click [download], it downloads. It doesn't try and get you to sign up.

Scribd looks like a big improvement on flash definitely though. The fact it's using fonts rather than fonts rendered on an image is lost on me though. Copy+Paste of text works fine in both viewers, and that's all that really matters to me.

We have discussed some really cool things to do with page thumbnails. It is one of the features I am particularly excited about, because I feel that navigation of online viewers has a long way to go. How do you like the slider/scrubber in terms of the nav experience?

For me, the slider thing duplicates the functionality of the scrollbar(browser generated), without adding anything else useful.

Also I'd much rather see the viewer separated from all the other clutter. Didn't it used to be?

We have some ideas about hooking the slider in with thumbnails. What we think the slider affords us is a nice way to give more context to moving around a document. Currently, it gives page numbers as you move and snaps to pages. These are outside the scope of the native scroll (to some extent) and we would like to take this part of the UX even further.

Try the fullscreen button -- is that what you're looking for?

Thanks yes.

Also the "readcast" popup is pretty annoying IMHO, popping up each time, obscuring the document.

If you have a Scribd account, visit http://www.scribd.com/account/edit#sharing and turn off all the Readcast options. The popup will be suppressed.

I don't have an account. I don't understand why I want an account yet :) The value proposition of what Scribd solves, is still puzzling to me... But thanks for the info.

FYI: Whenever I see "readcast" I read it as "Re-Ad-Cast" not "Read-Cast" as I'm sure you mean it. With the color on the page it's a different story, though.

Can't reply to you directly axod, so I'll put it here. To better support fullscreen viewing and other setups, we still have some behaviors to figure out with those pop-out dialogs.

I also agree that clicking download should let me download the file.

If you’re reading this, you’re using Scribd in HTML5.

Not so much, no..

I used to skip scribd search results and I won't anymore.

This is a million times better. Not just the reader (which is awesome) but the site design itself. Obviously a lot of thought has gone into it. I really like the way the footer/toolbar behaves when you reach the end of the page.

Well done!

This. Very much this. I used to avoid scribd because I didn't want to deal with the flash; now I'll probably favor it because I know I'll get a good layout.

This is a fine comment and I agree with you, but you might want to avoid the Reddit-speak around here. My guess is that's where the down-votes are from. Here's another case: http://news.ycombinator.com/item?id=1321795

Eh, I don't really care about downvotes. I don't usually bother to post if all I have to add is an agreement, but this was basically a request for opinions.

It does look really great, but more than that I wanted to thank you for a really classy post. I certainly thought I was in a timewarp when I saw the submission title.

Hey guys, I'm trying to understand Scribd's purpose as it's written.

Re: purpose 1) For publishing a body of text, or a hosting literal file type, there's already Posterous and, any free file hosting site.

Re: purpose 2) Discovering interesting content .. appears to be served pretty well by the likes of Digg, Reddit, Twitter-link-mining apps, etc. Sharing, there's many bookmarklets and "share this" buttons.

If the value is in unique content, but the content is provided by the crowd, then for an individual content creator, what's in it for them? E.g. if I have some really interesting content, I could post it on some public hosting site, or own my own web site, and if it genuinely is interesting, it'll hit the main page of a site like Digg/Reddit, and therefore sell itself.

I think the move away from plugin prison to HTML5 is great (am myself a developer), but we know that the mass majority don't buy based on techie cool-ness (if they did, then Apple's developer un-friendly-ness would drive them all to Android!)

Sounds to me like you guys want to be the YouTube for documents? (which IIRC wasn't that the goal in the beginning?) I guess like Smugmug's founder's interview, there'll always be a high-end niche of any market (are you trying conquer the "high-end" and "niche" documents? If so, then perhaps it might be a branding problem?) As in, if people perceive you as an expensive brand, then naturally people will _want_ to publish via Scribd. If I was writing an article about animals, I might be better off publishing the article on National Geographic's blog vs. my own no-name blog.

Other than that, awesome product! I guess I am still just trying to figure out why I should put my documents on Scribd.

Always a fan of YC-startups. Good work, guys.

These are all great questions - thanks for giving me an opportunity to clarify. Let me try to explain.

Regarding 1) We've done site surveys in the past to try to understand why people upload to Scribd. Every time we do it, we find that the #1 reason is: "Scribd gets me readers". Not "it hosts my docs for free", or "I like the display format", or whatever. Readers is what they want, and so we have worked very hard to maximize the number of readers we can get them.

I think file hosting sites are great, and Scribd is not attempting to replace them, they just serve a different purpose. If you want to share a file with a few friends, a file hosting site makes a lot of sense. But if you want to build an audience of strangers, posting on Scribd gives you a surprisingly good change of doing that.

If you don't believe me, try posting some doc you've written on Scribd. You will probably be surprised by the number of random people bumping into your document brought by related doc traffic, Scribd community traffic, search engines, Facebook, Twitter, our mobile apps, and all the other ways we make Scribd content discoverable.

re 2) Absolutely. Scribd isn't trying to be a content discovery site - other sites do a great job of that. Rather, if we've done our job right, then when you bump into interesting links on Digg and Reddit, some fraction of them will happen to be hosted on Scribd. Now that we've launched the HTML5 reader, you will hopefully enjoy clicking over to Scribd to read the document that Reddit surfaced for you.

Re making Scribd a "high-end" brand - I think that is a terrific idea. Creating that kind of brand is not easy, but it can be extremely valuable.

Interesting, so you really are becoming more YouTube-for-documents like.

I prefer to upload a video clip to YouTube and then embed it on my site vs. hosting it straight on my site, because if I upload it on YouTube, I get more visitors who may learn about my video (that didn't visit my site), and those may eventually be curious enough to visit my site.

So "discovery" is the big deal here, the "Scribd gets me [document] readers", like "YouTube get me [video] watchers".

Good to know :) And def +1 for doing your market research / cust. dev!

Thanks for the well-thought questions. Being as the web is all about content, there are many players and there is plenty of overlap. We do want to create a place where people want to publish their essays, poems, notes, etc.

I too am a developer, and it is sometimes easy to get caught up in the technological aspects of interesting projects such as this, our latest. I do feel, however, that we have/are creating a platform people want to use.

Here's an anecdote about some value I feel we began providing recently: I got to work on a little project we called Readcasting a few weeks ago. It basically allows people to passively or actively share what they are reading. A day or so ago, one of the ladies in marketing circulated an article[1] about how the social aspects of Scribd (to a large extent Readcasting) has helped an author gain a great deal of exposure. I thought that was pretty nice to hear. It's always a pleasure to see that people use and like the things you build, but it also pointed directly at the purpose of Scribd - to create a platform that benefits the people who publish on it.

1. http://www.huffingtonpost.com/kate-rados/where-marketing-mis...

You know what, you guys should collect these social proof points and feature them prominently on your first page. So that content creators visiting Scribd for the first time can immediately understand the value proposition right away. This link is a perfect proof point - and priceless, because for the sheer fact that it wasn't written by Scribd itself. You can write anything about yourself, but it's what other people write about you that matters.

Your brand is what other people say it is ;) You guys should leverage such proof points more in your marketing effort.

That's a great idea. We could definitely do a better job of surfacing stuff like that.

All up I like what you're doing but it's broken in Opera. After playing in Firefox:

- Text selection is really awkward, drag too far and you select unrelated stuff from nearby elements instead or as well

- The navigation stuff feels really weird down the bottom, probably because most software puts that stuff up the top

- The blue/gray page strip is weird, it feels more like a zoom tool rather than page navigation

- In book/slideshow mode the length of the page and the toolbar being all the way at the bottom makes me think I should still be scrolling

- Zooming in book mode has some of the site components overlaying it, in slideshow mode it adjusts properly

- Text selection is really awkward, drag too far and you select unrelated stuff from nearby elements instead or as well

Same in Chrome. Perhaps making images into CSS background images of the slides would fix this.

Thanks very much for the feedback! We'll investigate all of these. Opera support has become a high priority fix for us.

I've never had such a sudden swing in opinion about a product. Previously, I can remember real panic in seeing your logo pop up: this was about to be a waste of 20 seconds of browsing.

Now I'm reading PDFs on my iPad in an experience that very nearly beats some of the native apps.

You guys listened, pulled a herculean engineering job, and absolutely hit aces. 1 day into it I can't say how I'll use your site, but I'm now excited to find out. Fantastic job!

This is great. Kudos for having the guts to radically innovate while everybody else just complains.

A few notes:

The interface seems very cluttered. I know you have lots of exciting features, but if you want to make it about the content, then you need to find a way to hide this complexity. On my mind: http://uxhero.com/ux-theory/ipad-pages-vs-microsoft-office/

The right bar is really distracting and throws off the weight of the page. I really don't care what other random folks have read the document. Not really relevant until I'm tied in via FB.

I don't really understand why there are the different modes. Slideshows should be in slideshow mode; text should scroll. I still have to scroll with book mode, so what benefit does it offer?

That is great feedback. With an undertaking such as this, UX is definitely going to be a big issue. While we were careful and thoughtful in creating the current interface, we surely have plenty of work to do in terms of the overall experience. That's one of the aspects I am most excited about, though. We have so much more to work with - we will be able to make great strides in usability over the coming weeks and months.

That's a good screenshot comparison. I type documents in OpenOffice and have literally every toolbar disabled. If I need to change some text I use the right click menu or a hotkey.I also have my OS bar at the top of the screen to put the remaining clutter in the same area.

I want to see how it works on mathy latexed pdf documents. I've yet to experience an online reader that handles that robustly.

Check it out on this baby:


That's a LaTeX paper by one of the guys who built the reader.

you have just transformed my opinion of scribd's utility from negative to positive. kudos.

and by positive I mean awesome :-) the rendering is beautifully done

Really amazing work. However, when the scrolling to the images at the end of the paper, it kinda slows down my Firefox 3.6. Chrome 5.0.369.0 shows it smothly though (all on ubuntu 9.04)

Yeah can second this, managed to hang the browser after zooming in then trying to scroll on the latex based paper in FF 3.6, was a little slow, also in full screen seems to slightly hang at the page boundary.

It looks great in Firefox, but it's completely unusable in the latest Opera.

My Nexus One seems to be missing a number of symbols because I saw a few square boxes. Otherwise an amazing achievement - a very usable mobile PDF viewer.

Anyone on an iPad/iPhone/Pre able to see every symbol on that paper?

There's some text issues (text on top of other text) on my iPhone on the normal text fine. The forumlas look great.

It's pretty slow to render on my 2G (definitely rendering because I'm on wifi), but it is pretty underpowered and obsolete.

Font rendering issues on my HTC incredible. The navigation bar doesn't work either, it just hangs out all the way on the bottom of the screen.

If you get the chance, would you mind adding info about the issue you are having to this Google form? https://spreadsheets.google.com/viewform?formkey=dF9URVpxVEZ...

We are doing our best to document all of these and get fixes out as soon as possible.

Looks really nice. I zoomed in so I could read it, and then scrolled, noticed the nifty scroll meter at the bottom, and then it caused FF 3.6.3 to hang for quite a while (Win XP).

- when I change zoom mid-document, scroll to the same position in the text I was before

- modify default zoom level depending on the browser window size (I had to zoom twice before the size was reasonable)

- put the document-action keys on the top (the bottom is for statuses). It took me long enough to realize the bottom pane was where things were that I would have left if this were a paper I came across on HN/prog.reddit

- don't force registration to download, that's just rude

in my browser (ff 3.5.9) fonts are very small $(

Try clicking the 'zoom in' button in the bottom bar.

It should probably be zoomed in more by default. (2 clicks more seemed right to me -- safari on a 13' macbook)

does fullscreen help? what's your screen resolution?

That's very impressive. Any idea how these guys achieved the rendering of all those math symbols so flawlessly?

Nice at initial size, but zooming jumbles the words in the latest Chrome Mac.

Holy shit that's awesome. I've longed for the day that latex2html dies a well deserved death and there aren't ugly little rasterized images where every mathematical symbol goes.

I personally would like the pages of the document scroll horizontal rather than vertical. What would be really cool is identifying the columns in a pdf document and formatting them such that it shows two or three narrow columns that I can scroll through. Basically something similar to this (http://amarsagoo.info/tofu/) on the web. This would make reading on the web a lot more easier than it is today.

That's a really interesting idea. Given our other view modes, I don't see why we could not do a horizontal scroll as well. I'll keep this in mind.

I just saw the new HTML docs for the first time using Chrome on Linux and it blew my mind. It needs a bit of polish still it seems on some of the font anti-aliasing but overall it's amazing. Lightweight, easy to scroll, fast and oh - look... my CPU usage isn't spiking to 70% because no flash is running :)

Great work guys - you really have upped the ante and I hope given Flash the deathstroke by proving it's doable.

I'd recommend if possible that you provide some of your tools and code to the web as open source - if you can make it easier for other people to dump Flash you could truly start a revolution worth joining.

We're looking into releasing as much of our HTML5 conversion as possible as open source. For starters, basic technical details on how our conversion works will appear on our tech blog http://coding.scribd.com in the next few days.

Awesome! I don't think even Google did that.

A number of our engineers, particularly those who made this all possible, will be adding posts to our coding blog in the coming weeks. They should help kickstart any effort to render as we have.


Nice, I would be particularly interested in the conversion from PDF to HTML, and how you handle the word-spacing and line breaking. I've been doing some work on rendering TeX-like data in HTML (http://www.bramstein.com/projects/typeset/) and it would be great to see how you are solving similar problems (assuming you're willing to share these kind of details.)

I'm not sure if this is happening to anyone else, but I'm thinking the viewmode default should be on 'slideshow'?

Otherwise looks great! I can't wait to see it formally launch. Is there a way to upload documents in different formats now to test it out?

I literally just changed this off of that view mode! We have a number of documents that we are launching this with, and we wanted to showcase a number of different view modes.

LOL, your usernames are different from the original poster, but I'm guessing a few fellows from scribd HQ are monitoring this thread. =)

Thanks for the responses! I'm assuming then that the default viewmode will change depending on the document (ex. an e-book will default to 'book', and a pptx doc will default to slideshow)?

I will also 2nd the request for keyboard shortcuts as suggested below. Maybe activate page-up & page-down in full screen mode.

As also mentioned already, page thumbnails would be great.

Yes we will do some detection based on document type for the default view mode.

With regards to default viewmodes, that's working for iPaper and is coming soon for the HTML5 viewer.

Go ahead and upload documents in any formats supported by Scribd ... they'll be added to our conversion queue and appear as HTML 5 documents when ready.

About 9 months ago one of the Scribd cofounders emailed me to follow up on comments I made on HN tearing his product to shreds, and while he was surprised that I had detailed answers to his questions, I think I was even more surprised that he was up for a long involved conversation about the problems, given the sorry state of his product. I got the impression at the time that they were focusing on the site and community to the detriment of making the content actually readable.

Now the product is finally living up to the founders :)

That conversation was one of the pieces that led to this change. Thanks for taking the time to talk to me about the product.

I'm very glad to see you guys moving in this direction. I, frankly, hated Scribd before, even though you guys had a relatively nice PDF reader.

The only suggestion I have right now is to have the full screen button take into account the longest side of the page and the browser window size will scaling. I think it would be more useful to have the shortest side of the window be the limiting factor with the scaling, rather than scaling to the longest side of the window. That way, content isn't hidden.

Selecting text is a bit odd for me- on slide 4's bottom text, for example, selecting a large part of it selects the whole slide. (Chrome w/ beta channel)

We're aware that there are text-selection inconsistencies (caused by the way browsers handle text above images). We're working hard on improving that.

There are some oddities with selections because of layering, but this is something I often notice in native PDF viewers. Have you seen any particular patterns which cause this?

The only pattern that I can recognize is that it's impossible to select an entire line character-by-character with the mouse- I can do it by triple-clicking the line, though.

That's helpful, thanks!

I'd like to see larger up/down scrolling buttons. Fitt's Law is working against me when I try to click on those. I think they should be larger because it's the controls that users are most likely to use. Maybe Marco Arment's Instapaper pagination will give you some ideas: http://blog.instapaper.com/post/414438490

Overall, a huge improvement. Good work!

There is something very, very wrong with your search. Doing a search for 'Lisp' yields 22 pages all containing the same book 'On Lisp'. I tried doing an advanced search on titles only and got the same result.

On a somewhat related note. I find it odd that you don't have a specific topic or category for Comp Sci/Programming. Just a topic for 'Science and Technology'. What is the difference between a category and a topic?

Nice. Any word on when we'll be able to use this in conjunction with the API?

I just pushed a new build of Twiddla that allows viewing of the new Scribd HTML5 docs. Anybody feel like doing a bit of testing for us?

  - go to http://www.twiddla.com/1
  - browse one of these urls:

How are you handling the legal issues around the fonts? My understanding of most foundries' rules is that pdf/flash embedding is ok but eot/otf embedding is not. I can see by playing with font families in firebug that you're subsetting heavily on your html5 example. is that all that's necessary?

Old Scribd used to wrap open-format documents in a proprietary, platform-limited viewer. Old Scribd was evil. New Scribd takes documents and makes them even more accessible and open. New Scribd is good.

It will take awhile to get the bad taste out of my mouth, but this is a huge step in the right direction.

I would suggest typing the up and down arrows at the bottom to two keys, j and k would be appropriate. Then have the mouse-overs for those buttons include the keyboard shortcuts. Yes, you can accomplish the same thing with spacebar and arrow keys, but it is annoying to jump to a spot halfway through a slide then have to scroll.

The functionality is already there. But I personally don't like to have to aim my mouse to navigate when I'm reading a long document.

Another idea to consider is creating an iframe version of the page, which would allow you something closer to the old experience. The nice thing about that is to allow third party websites to transparently embed scribd documents in their pages.

Other than that it looks good and is more pleasant for me than your flash interface.

We totally love your j/k keybinding idea. Consider it implemented. Maybe we'll additionally do s/w, a binding that comes more naturally to people unfamiliar with VI, but familiar with computer games :)

+1 for J/K. J and K for navigation is becoming quite prevalent, as seen in websites as varied as Google Reader and The Big Picture.

Gmail uses these keybindings too :)

Curiously, the up and down arrows threw me off. I expected left and right buttons.

Good job! Bits of feedback:

'Back' doesn't work. When I got to the bottom of the intro link, and clicked the upload button (just because I could), and then hit 'back', it brought me back to the top of the document instead of where I left off, which breaks normal browser behavior. I'm not sure what's causing this, because when I hit back, I do see the page counter at the bottom show 20/22 briefly before getting reset to 1.

Others have already mentioned not putting the navigation at the bottom so it doesn't look like spam.

Is an Arc90 readability-like feature in the works for those of us who want to read and not fight the designer who thinks grey-on-slightly-darker-grey with a hard-to-read font is edgy? (Site is currently broken when trying to use readability.)

Hmm, that first problem should not be the case. I'm unable to reproduce it in my browser, but we will look into it.

What exactly do you mean by the nav at the bottom looking like spam? I assume you are referring to the sticky toolbar?

Also, if you would like, we are doing our best to keep track of issues people are seeing. You are welcome to add your issues to this Google form: https://spreadsheets.google.com/viewform?formkey=dF9URVpxVEZ...

Filed! ('back' works elsewhere, was broken on Chrome 4, winxp) You might want to use something that will capture the user agent from the browser itself as well as asking the user about what browser they are using. AFAIK, Google docs does not let you do this.

The sticky toolbar looks like spam seen on other sites, esp because of the Readcast FB/Twitter share bubble that pops-up. (Not quite as obnoxious as the double-green line underline, but still.) In addition, the header that you do see when you first load the page draws attention away from the bottom navigation.

Great, thanks for filing that. We will look into better presentations of the bottom toolbar.

Beautiful! Great app, guys.

The things that stands out to me right now, both good and bad, are

1) how much faster it loads compared to Google's reader

2) how much laggier it is when scrolling compared to Google's reader

3) When I land on the page, I look to the right and see half the page taken up with meta-content (ratings etc) – it's pretty dominant. Perhaps there's a way to tone it down a bit?

4) The selection issues other have brought up

5) It's beautiful!

6) I'll be using you guys a lot more (reading, and eventual publishing) now that you've done this :)

(Used the link axod provided: http://tinyurl.com/33339j8)

As for #3. Try our fullscreen mode. It's awesome.

about 3: there is a way to tone that down- click on "fullscreen" (the square with the arrow) on the bottom toolbar.

Wow. That is gorgeous. You've just made me really excited for the future of the web.

This is seriously impressive stuff. It's wonderful to be able to right-click and run 'Inspect Element' on everything. Of course that's not the motivation for doing it (you laid out the real reasons well), but it helps in winning over geeks like me.

Also kudos on executing things in the right order: start with commodity tech but build out innovative and useful features. Then rewrite the underlying tech to make leaps in the level of refinement.

I like your concept, but I'm sour on your company because of the blatant piracy that's allowed. I know filtering content is hard, but I think you could do a better job.

For example:


You can replace Hackers and Painters with most any book, and it will come up on Scribd.

Just my $.02.

As several other commenters, this change actually makes scribd usable to me (flash is not stable enough to run the scribd viewer on my linux box).

My remaining problem with scribd is that there are documents that I would like to buy (recipes), but since I live outside of USA, that doesn't work. Would Paypal really be so hard to use for international payments? Or is it a problem with regulations?

Thanks! Yes, Scribd never worked well under linux before.

Unfortunately, the international payment restriction is totally a legal issue, not a technical one. What country do you live in? We are finding ways around the legal issues country by country.

I live in Sweden. Unless there is some simplifying EU framework, I guess that Sweden is not that high on the list (which is understandable).

On a separate note: the new reader looks great. One comment; it;s a bit wide for netbook screens. I appreciate it is not possible to support all screen sizes but us netbook users are a reasonable share now :) The page content fits but the tools/info is off to the right.

In fact everything on this screen appears "larger than life" (Chrome Dev channel on Windows XP, Samsung NC10)

Excellent point about the netbook screens. Thanks!

I've spoken with a few lawyers who would love for Scribd to allow deep links into PDFs/content that would bring a user to a specified page and highlight a certain range in a document. Think "document permalink" so that they can send the URL to their colleagues.

Also, commenting and annotating on the doc. i.e. highlight a paragraph and add a line of text in a bubble.

That would be neat, kind of like GitHub's links to a specific line or ranges of lines.

We've talked about doing both of those things. They're high on my to-do list.

Some text in the presentation is not browser-searchable (e.g. looking for "navigate" in Chrome does not work for me)

I can't tell you how great the experience was! The one minor UX issue (and its more of a feeling than anything else) is the controls at the bottom. It just kinda feels like I can't see something. maybe the actual controls are too close to the bottom of the page (maybe its my monitor), who knows.

But the experience was excellent!

Really glad you are enjoying it. Are you talking about the toolbar auto-hiding? That was a decision we went back and forth on, so I would imagine it is not set in stone. Would you prefer it expanded at all times?

i really disliked the way it auto-hid. I'd rather it remain visible.

Agreed. Maybe instead of auto hide, always leave it on with an option to hide it.

We'll consider making that a preference, so you can prevent it from hiding.

guys, this looks awesome. it's obvious that a ton of work went into it. congrats!

i really like the change of direction, too -- it's becoming more and more obvious that betting your company on flash is not a good long-term strategic move.

but if nothing else, i'm happy that it makes my reading experience so much better :)

One thing I've never got to the bottom of is: do the vacuum'd in documents (link the links added to HN submissions) get stored/archived and made accessible on Scribd?

That's never felt a very comfortable thing to me; though I admit my main awkwardness is in how it is automatically appended here.

All documents uploaded to Scribd (either automatically or manually) are stored in S3 and are available on Scribd until the document is deleted from Scribd. If you find your content was uploaded to Scribd without your permission, you should contact our support team; they'll take care of it for you.

Does it appear in document lists? Or is knowing the URL the only way in?

The "ethical" concern I have is that it encourages behaviour like that shown on hn - with the scrape link added automatically. There is a distinct difference there compared to people uploading documents themselves (IMO anyway).

Of course if it's stored but not indexed that feels fine.

(thanks for clarifying)

HN is using the "slurp" API we make available at scribd.com/developers. Documents uploaded via the slurping system are marked as private and uploaded to a fixed "slurp" account for which no one (outside of Scribd) has the password.

Nice post. I've always thought that scribd provided a service for people who wanted to share correctly formatted documents, even if it had the downside of being locked into Flash it was better than nothing. The HTML5 reader is a lot nicer.

I'm still planning on copying all your social features.

I think if you could do a left-right layout or ajax one at a time for the document pages, then it would feel more like reading a book and increase the likelihood that users will notice the bottom tool bar and use it for navigation instead of the side scroll.

I found this a terrific reading experience. I love the fact that everything just appears in one page.

It's a huge achievement to pull this off, preserving the fidelity with the original text. Many congrats to Scribd for pushing the state of document-sharing forward like this.

This is something we are quite proud of. You will get to read more about how it all came together when some of our engineers like matthiaskramm publish their blog posts.

Awesome upgrade. I've always wanted to read Scribd documents on my phone, but had to do the bass-ackwards method of downloading to PDF then uploading to my phone.

Are my existing docs in HTML5? What happens now to embeds?

We're in the process of back-converting existing documents to HTML5. Most of your recent public documents should be converted now or soon, with private docs to come later.

I just tried the site and it's broken on WebOS. The thing I was looking forward to the most was being able to use Scribd on my Pre, so that's what would make Scribd the best reading experience for me.

Totally hear you. For the first version we focused on making it work on browsers that support @font-face. The next version will be all about making it work pretty well (if not perfectly) on browsers that don't.

That's a great point, though, thanks for letting us know how important that is to you.

insanely great.

In FireFox 3.0.19, it fails to render images or fonts, throws two javascript exceptions, then pegs the CPU at 100% until you kill the process.

Nice in Chrome though.

Whoa. That's .... not good. What OS? And you're viewing the http://www.scribd.com/html5 document?

Suggestion: left/right arrow keys should map to the previous/next page buttons.

edit: it does in slideshow mode but not the other modes.

The font in that document is really hard to read at times. Pixelated and blurry. And the little tool bar at the bottom popping up all the the time is annoying.

That said, I applaud the effort and see it as a HUGE step forward from the old flash version.

Use the zoom.

First of all, the HTML5 documents look great.

However, how do you deal with font licensing? You are using @font-face, but this is still not permitted for many commercial fonts, even if they allow embedding in PDF.

Wow, this is seriously awesome. I always avoided Scribd because I could never just download the PDF, but with HTML5 view + fullscreen mode I am happy. Now time to overtake Slideshare... ;P

Minor point, do the "copy" buttons really need to be flash still?

Actually, they do. It is unfortunate, but flash is the best way to access the clipboard from the browser. There are some slick implementations that make the use of flash less evident, but if you use something like flashblock, you will surely know. See: http://github.com/mojombo/clippy

Just a suggestion - say if I skip ahead to the 20'th page, then Scribd will load all 19 pages and the 20'th page.

That might not really be necessary and it does make the loading slower.

This behaviour shouldn't be happening. Which browser are you using? How are you skipping pages?

chromium on linux - I skip ahead by clicking the small dashed bar (which has as many dashes as pages)

Thanks. We'll definitely look into this.

this is really impressive -- high fidelity, search, scrolling, no waiting for flash to load. congrats guys! we'd love to get this kind of previewing on dropbox

For those curious about how to pronounce Scribd (as I was for a long time), it rhymes with "ribbed", as in, "Scribd for her pleasure."

Amazing work.

Accessing the site with Javascript disabled, I click Download and nothing happens.

So you basically want to make scribd.com some sort of PDF aggregator?

It was already a PDF/DOC/etc. aggregator -- that's always been their entire business model.

What's new is that they're demonstrating that they care about people actually being able to read the content.

Nice work guys!

Is there a vacuum URL for HTML5?

I can't view docs on iPad?

What problems are you having? We did tests on an iPad and created a rather sleek, lightweight version of the viewer for it.

I can see a list of docs, but can't seem to open them (the ipad opens the pdf just fine). I thought clicking on the thumbnail or doc name would open it.

Nope. There's the download button. There's the thumb and title. There doesn't seem to be a way for me to view the doc, short of clicking the download button for the PDF (which works). Actually, why go through the trouble of htmlizing the doc when i can just view the PDF (which renders beautifully on iPad)?


Um, google was already indexing all documents on scribd. And while it's possible this will result in better SEO, I think it's more likely that all the html necessary to build documents with layout will confuse google's crawler/parser rather than help it. So on balance, I'd bet this is at best neutral for SEO and at worst pretty hard on it. I think Scribd really is doing this for the reading experience.

The #1 problem I have with scribd is that you appear to be unapologetic content thieves.

Take, as an example, the [scribd] links that HN generates automatically. That's your earliest investor automatically stealing and rehosting content in violation of both your TOS and the law, and you do nothing about it.

You must be aware of it, but you don't stop it. That certainly gives the appearance that you don't care that you make a large portion of your money by screwing over rightsholders.

Truly, I believe the world is a worse place because your business exists, and that if you shuttered your business today the gains would vastly outweigh the losses experienced by all. I understand that the DMCA allows you to operate with impunity, but legal doesn't mean ethical. In this case, it's just abuse of a loophole in a poorly written law.

This is my entirely honest opinion. I continue to hope you'll close down scribd, or at a minimum engage in a MASSIVE rethinking of how you deal with copyrighted content. I know this won't be a popular opinion, but I simply can't understand how [scribd] links are anything other than a tacit admission that your real business is the monetization of pirated documents.

Looks gorgeous. I'd make the tool tray fall down a bit faster, but it's awesome. And I love the subtle blue lights across the bottom.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact