Amanda Hickman

Starting from scratch, where would you start?

Last Friday was my last day as program director of DocumentCloud, a catalog of primary source documents and a tool for annotating, organizing and publishing them on the web. I've got a few more talks lined up (I'll be showing off DocumentCloud at SEJ in Miami and talking about our work at MobilityShifts), and I'll still be part of DocumentCloud's advisory group, but it was time to hand the reins over to the (very capable) staff at IRE. For my next trick, I'm helping get a new accountability journalism project off the ground.

I've been thinking, as I navigate this new project, about a couple of good questions.

Real Collaboration

One is this: What does an honest collaboration between bloggers and reporters look like? Local bloggers are a great source of story leads, but they deserve at least a nod of gratitude when their footwork yields a real report. Take a look at Miss Heather's roundup on New York Shitty of unattributed reporting if you don't believe me.

I'm not convinced that traditional media is refusing to enter the link economy. I suspect that traditional media is intensely competitive and the creative legwork of reading some blogs doesn't get you the credibility in the edit meeting that pulling a story out of thin air gets you. But I don't know that.

I do know that neighborhood blogs are a great resource, spot some great stories, and often don't dig past a post or two. So they're ripe grounds for finding leads worth following up on. Giving local blogs credit when you take a lead of theirs and run with it is a great first step, but could we do more?

Starting from Scratch

Another is this: We're starting from scratch. Where do we find the balance between building the perfect content management system and overbuilding? I don't think the CMS in general is broken, outright, but the news CMS? Broken.

Erik Hinton, TPM's technical fellow, has been writing about the problem and taunting us with tales of Baroque, which not just anyone can use. (Or is Baroque available in the wild?)

Journalism is different from blogging, but most days we're trying to squeeze a news CMS out of blogging software. News organizations have data to map and documents to annotate. Every story has context, and when we hit our stride, old stories will start to have followups and new stories will have histories. Our work will be reproduced, reprinted and translated, and we want readers to be able to find those conversations and read those translations. Stories are updated and corrected: We'd like to be responsible and show those changes. As our archives grow, we want people to be able to navigate them without frustration.

So I'm wondering, if you were starting your newsroom's website from scratch today, how would you do it?

Image courtesy of flickr user mujalifah.

I hopped down to New Orleans this week, to tell even more journalists about DocumentCloud, a catalog of primary source documents and a tool for annotating, organizing and publishing them on the web, and had the opportunity to sit down with The Lens' Ariella Cohen and Steve Beatty.

Steve is doing some great work on a charter school reporting project, covering every New Orleans parish school board and incorporating documents about many of them. Most of what we discussed is what I talk to most journalists about: how they approach their work, what they see as their mission, and the reporting and community information landscape in the places they cover.

But the conversation veered toward their DocumentCloud pet peeves, including the way that our file uploader automatically strips dashes and underscores from file names to propose a document title, even if those dashes are separating the month from the day from the year in a date in the file name.

We can fix that (I'm pretty sure we already have), but it reminded me once again that with a project like DocumentCloud (and there are a few such projects in this round of Knight News Challenge winners) there's no substitute for getting out and talking to people long enough to hear what really bothers them. The smallest things make a big difference in usability.

P.S. We really love this example of an embedded annotation from the Center for Public Integrity. The signatures are part of a longer report on payday lending that you can find here and here.

As DocumentCloud settles into our comfy, new home in Missouri, we're quite pleased to welcome Ted Han aboard as our new lead developer. Though our prior lead developer, Jeremy Ashkenas, has moved to a full-time position at The New York Times, he continues to be an active and enthusiastic contributor to DocumentCloud's open-source tools and platform. DocumentCloud is a catalog of primary source documents and a tool for annotating, organizing and publishing them on the web.

Ted joins DocumentCloud from Videojuicer, an online video platform focused on open standards and software integration. He's a computational linguist by degree, developer by trade, and Sci-Fi nerd by leisure. He was selected as a participant in the Knight Mozilla Journalism Challenge and has worked on DataMapper, Merb and a variety of data-based projects, including the CrisisCommons response to the Tohoku earthquake & tsunami.

We've made plenty more changes to our staffing this summer -- IRE's Resource Center director Lauren Grandestaff is fielding new account requests, and more and more support replies will be coming from IRE staff this fall. I'm still spending some of my week helping everyone get their bearings, but I'll be retreating to an advisory role soon as well.

Ted is planning to get right to work on DocumentCloud's planned reader annotation tools. His first task will be to identify the newsrooms that DocumentCloud will be collaborating with as we build out those tools, so if you've been yearning to try out calling on your readers to provide annotations on a document, now would be a good time to start talking to your editors and web producers about whether this is something your newsroom should be helping us test.

Ted (who blogs here) is new to Columbia, Mo., and I know he has some big ideas about kickstarting a Hacks/Hackers chapter, so we're thrilled he's agreed to join us.

When President Obama released his birth certificate and dozens of news organizations turned to DocumentCloud to present it to their readers, I snarked a bit. Though the birth certificate did prompt a few questions -- which we're still navigating -- about the best way to handle duplicate uploads, the secret truth was we were both proud and flattered that so many newsrooms, faced with a document they wanted their readers to see, came straight to us and knew they could count on us.

When half a dozen newsrooms turned to us to help them get Sarah Palin's emails out to the public, our first thought was that we'd waited too long to address the duplicate documents challenge. Our second was that this kind of case -- multiple newsrooms eager to scan the same tower of documents -- is exactly what makes managing duplicate documents challenging. DocumentCloud is a catalog of primary source documents and a tool for annotating, organizing and publishing them on the web.

Most newsrooms downloaded a PDF of the president's birth certificate and then uploaded that file to DocumentCloud. A simple "chksum" could have identified those as the identical files. It takes human intervention (or at least more complex algorithms) to declare that one scanner's photo of a particular page matches another's.

STOPGAP IN A BROKEN SYSTEM

The Palin email release was also an unfortunate reminder that DocumentCloud is often a stopgap in a broken system: There's no good technical reason those emails had to be released on paper stacked into file boxes in order to be reliably redacted. There's no good reason that a half-dozen newsrooms were sending staff to Juneau, Alaska, to digitize documents that started out digital.

Our own system was feeling a bit broken as well -- we knew that reporters from multiple national papers were assembling in Juneau and preparing to be the first to get the documents online, but we couldn't convince any one paper to scan the 25,000 pages once and make them available to everyone to embed. That left us spending a few hundred dollars to queue up additional servers so that our users could duplicate each other's work. Not too much money, in the grand scheme of things, but money nonetheless. Those 25,000 pages turned out not to contain much noteworthy information, but that's a different story.

We were also queuing up to make a big announcement of our own, the merger of DocumentCloud with IRE. I was too busy keeping our own shifting sands level to compose any particularly coherent thoughts, though I was proud of the way our servers weathered the storm -- DocumentCloud was slower than usual for a few days, but not by much.

We designed DocumentCloud to scale as necessary, and it scaled well. And there's no question that once again we took it as the sincerest of compliments that the country's top newsrooms turned to DocumentCloud as soon as they had documents to share with their readers. I can't think of better evidence that DocumentCloud has been a success.

FACILITATING CIVIC CONVERSATIONS

Earlier this week I found myself discussing Internet activist Aaron Swartz's indictment. I kept seeing tweets like "downloading too many articles is a felony?!" Some of the charges in the indictment, like breaking into a locked room, go a few steps beyond coming up with a workaround to avoid paying for your 21st New York Times article online. Someone pointed out that particular charge was bizarre because in the case of the locked server closet, MIT ought to be the aggrieved party -- these charges aren't coming from MIT.

All over Twitter people were having the same conversation and a decent number of other ones. The same was happening over email and in social networks. Over and over, all kinds of people were pointing back to the indictment. Hundreds of thousands of people read the indictment. And that is why we built DocumentCloud -- to facilitate civic conversations, to give people a better way to talk about substantive issues in the news.

Aaron's legal troubles are the kind that my own community of programmers, journalists, web watchers and advocates are especially interested in, so I've seen how having the document available has influenced these conversations, but I also know that DocumentCloud is full of documents like Aaron's indictment -- documents that journalists thought were worth sharing and that readers are skimming to measure against the resulting reporting.

Photo of Aaron Swartz courtesy of Fred Benenson and used here under Creative Commons license.

DocumentCloud is beyond delighted to announce we've found a long-term home for our project. We're merging our operation with Investigative Reporters and Editors, a non-profit grassroots organization committed to fostering excellence in investigative journalism.

This transition means that DocumentCloud, a catalog of primary source documents and a tool for annotating, organizing and publishing them on the web, will have a permanent place in a longstanding resource for investigative reporting. IRE has a long and established history of supporting investigative reporting, and we'll be a proud part of their ongoing work to provide journalists with tools that support their reporting. It goes without saying that DocumentCloud is a natural fit for an organization that has been upholding high professional standards and instilling a passion for public service journalism for more than 35 years.

IRE will continue to honor all of the promises we have made to our users, and our staff will be working to ensure a smooth transition. The best way to get your questions answered will still be reaching out to support@documentcloud.org or contacting us through the workspace. We're still welcoming new users -- contact us to find out more about bringing your newsroom onboard.

We've even got some great new tools in the works. More on that soon.

Moving to Missouri

All of us are committed to the continuing success of DocumentCloud. Over the next few months, we'll be handing off day-to-day responsibility for managing DocumentCloud to IRE's staff based at the University of Missouri in Columbia, Mo. I'll stay on as program director through the summer to facilitate a smooth transition. Developer Sam Clay is moving to San Francisco to join a startup there. Our lead developer, Jeremy Ashkenas, has moved to The New York Times' Interactive News team, but will remain actively involved with DocumentCloud on the technical side.

Our founders will be staying on as advisers to help DocumentCloud continue to thrive -- Scott Klein, Aron Pilhofer and Eric Umansky will remain on the project as advisers and advocates. We're already interviewing strong candidates to take over as lead developer, but will be looking for more developers, too. More on that soon as well.

DocumentCloud was first envisioned by a team of editors at ProPublica and The New York Times, and was founded in 2009 through a grant from the John S. and James L. Knight Foundation to build an online catalog of primary source documents and a set of tools to help journalists get more out of source documents. We are all immensely grateful to Knight for their confidence in us. We think their investment paid off. Not only do newsrooms have a new resource that is already indispensable, but DocumentCloud helped demonstrate that 21st century newsrooms are ready to collaborate and share what were once privately held materials. The public is better informed because of it.

Since we launched in March of 2010, newsrooms and watchdog organizations have used DocumentCloud to analyze, annotate, and publish thousands of documents ranging from suspicious, if not outright spurious, expense reports filed by local authorities in Long Island, N.Y., to hundreds of pages of correspondence released by the Financial Crisis Inquiry Commission, and much, much more. How much more? We encourage you to search our public catalog and see for yourself.

Embedding notes makes it even easier for reporters to bring source documents right into the story. One of DocumentCloud's primary goals is to make it simpler for news organizations to show their work -- to invite readers to review the very same documents the journalists used to draw the conclusions in their reporting.

The latest addition to our toolbox, which we quietly rolled out last month, allows reporters to embed a single annotation in a story online, and we've been delighted to see newsrooms making excellent use of it.

City officials in Torrance, Calif., circulated a press release explaining their decision to exclude some categories of crimes from the city's online crime map. The Los Angeles Times, whose prior reporting had raised the issue of missing crimes, posted a brief item online in which they embedded the annotated press release directly. Not only were they able to clearly distinguish the portion of their post that cited the city's words, their readers had the opportunity to review the city's complete press release on the subject and discuss it.

Meanwhile, when allegations surfaced that Immigration and Customs Enforcement intentionally misled local authorities about whether or not they would be able to opt out of a controversial enforcement program, Mother Jones shared 17 pages of correspondence between the legislator who raised the concern and immigration officials. In their reporting, they were able to highlight the particular passage that raised red flags for legislators. Mother Jones readers had the opportunity to decide for themselves whether or not the particular passage was vague, damning, or being blown way out of proportion.

Beyond embedding notes, we've been working hard on the little stuff, too -- improving our redaction tools, clearing up quirks in the annotation process, and giving everyone new ways to search our catalog.

What annotations would you like to see embedded?

As we watched traffic stats skyrocket last month as newsroom after newsroom uploaded President Obama's birth certificate to DocumentCloud and then embedded it, my reaction was hardly one of joy.

Why on Earth is a birth certificate more interesting than, say, the pages and pages of receipts documenting some outrageous meals (15 steaks, two orders of fish and a lamb chop -- for five people submitted by National Grid to the Long Island Power Authority after their Hurricane Earl cleanup)?

I like to think these are the documents we built DocumentCloud for -- that we're here to give a leg up to reporters scrutinizing spurious spending reports (reporting that prompted a formal state investigation) or documenting patent dishonesty and the unusual lengths one California town went to in order to conceal extraordinary salaries paid to city officials.

Vote of Confidence

Forgive me if I was underwhelmed by all the attention that the birth certificate got. My esteemed colleagues, however, helped me see the bright side of the flurry. For one thing, it was fast. Within minutes, 10 different newsrooms had uploaded the birth certificate and embedded it.

That says a lot: It says that when they have something they know their readers want to see, reporters turn to DocumentCloud. That's a huge vote of confidence in us. Plus, we didn't falter under the weight of the tenfold increase in traffic -- that's solid architecture for you. We built DocumentCloud with the hope that we could improve the way newsrooms share source documents with their readers, and at that, we're thrilled to be succeeding.

Increasingly, DocumentCloud is a resource for breaking news. When the news broke that Osama bin Laden had been killed in a town called Abbottabad, a search for "Abbottabad" turned up some pretty rich stuff, most notably that a former Gitmo detainee led U.S. authorities to the Pakistani town back in 2008.

New Feature Roundup

Meanwhile, we're still listening to our users and looking for more ways to make DocumentCloud easier to use and to help reporters give their readers the documents behind the story.

We're looking forward to seeing what our users do with our new tool that lets you embed a single annotation, and we're excited to watch the great uses newsrooms have put document sets to.

From embedding documents accumulated over two decades spent covering an Oregon commune where things went horribly awry to sharing the documents detailing the Federal Reserve's support for ailing financial institutions, or the background material from coverage of a profoundly embarrassed local philanthropist, reporters seem to be getting the hang of embedding document sets.

So we have a question for the reporters who have been using DocumentCloud already: What would have made this even easier for you?

Investigative Reporters and Editors (IRE) announced their medal winners this week, and we were impressed to see that both winners wove DocumentCloud into their winning reporting. Since 1979, IRE has honored outstanding investigative work with their annual awards. This year they honored a Los Angeles Times series on outrageous salaries in one of California's poorest towns and a collaboration between International Consortium of Investigative Journalists and the BBC for a report on the global asbestos trade.

Breach of Faith

Los Angeles Times was awarded an IRE Medal for Breach of Faith. An investigation of financial impropriety in a small town revealed what turned out to be a dramatic case of corruption and mismanagement in the quiet city of Bell, California. Los Angeles Times reporters uncovered exorbitant city salaries (including compensation packages topping the million dollar mark) and errors in financial reporting that were more than just
mistakes.

They used DocumentCloud to post the falsified salary information that city administrators had provided to concerned citizens years earlier, and the subsequent indictment of the administrators who provided those false records.

Dangers in the Dust

International Consortium of Investigative Journalists and the BBC share an IRE Medal for Dangers in the Dust: Inside the Global Asbestos Trade. Throughout their year-long reporting project, they added all manner of document source material to a growing archive of documents.

Great work, and congratulations to both teams.

More Prize-Winning Work

We've seen other prestigious journalism awards go to DocumentCloud users, too. Last month, Alex Richards and Marshall Allen (who has moved from the Las Vegas Sun to ProPublica since) were honored with the Goldsmith Prize for Investigative Reporting, given each year by Harvard's Shorenstein Center to honor investigative reporting that "promotes more effective and ethical conduct of government, the making of public policy, or the practice of politics" for their in-depth report on hospital care in Las Vegas.

Alongside each of the five stories that made up that award-winning coverage, Richards and Allen used DocumentCloud to share their source documents with readers. It's great reporting and exactly the kind of work we imagined we could help support when we set about building DocumentCloud.

At least one finalist for the Goldsmith prize also put DocumentCloud to excellent use. ProPublica, NPR's "Planet Money" and Chicago Public Radio's "This American Life" collaborated on Betting Against the American Dream, an alarming expose of Wall Street's role in exacerbating their own meltdown. ProPublic used DocumentCloud to detail correspondence with their uncooperative subjects.

We quietly opened DocumentCloud's catalog to public searches in January, and we've been working since to do more with the great documents that reporters have added to our catalog.

When Vancouver Sun investigative reporter Chad Skelton asked if there was a way to automate display of the growing cache of documents he was retrieving from the city's ferry authority, the best answer we could offer was to point his readers to a search for the DocumentCloud project he was stashing them in. Our goal from the outset has been to help news organizations make their own substantive reporting more engaging online, not to drive traffic to DocumentCloud.org. Moreover, Chad was far from the only reporter asking us to make it easier to embed whole document sets. Homicide Watch even built a JavaScript widget to embed their sets. So the latest DocumentCloud feature, built out by our own Samuel Clay, is embeddable document sets.

Any DocumentCloud user can embed pretty much any set of documents on their site. It works whether or not the user's own newsroom originally published the documents. This means that the Vancouver Sun can embed their ferry documents, and that any user can embed a set of documents matching a search for Scientology. Documents initially published by the New Yorker will open on newyorker.com while documents that were published by ProPublica will open there. Someone could also embed the complete set of public documents that match a search for former Illinois governor Rod Blagojevich:

More Tools

We've added plenty more tools to help newsrooms get the most out of DocumentCloud, too. A dozen different "How do I ..." questions led us to dramatically increase the options available when users publish documents. Plus, a brainstorming session with American Public Media's Andrew Haeg in the halls of this year's Online News Association conference led to a tool newsrooms can use to share documents with reviewers outside of the newsroom.

Our users continue to help us make the most of the tools we've built, too. It's been a few weeks since the unstoppable Chicago Tribune news apps team released dcupload, but the python script, written against our API, makes it a whole lot easier for DocumentCloud users to upload a great heap of documents in one fell swoop.

DocumentCloud's Jeremy Ashkenas collaborated on this post.

It has been less than a year since DocumentCloud began adding users to our beta. Late Monday morning, a user uploaded our millionth page of primary source documents.

The thousands of documents in our catalog have arrived in small batches: five pages here, twenty there. The vast majority of the 65,000 documents that those million pages comprise remain private, but we're fast closing in on 10,000 public documents in our catalog.

Broad Appeal

Journalists are using DocumentCloud to publish all sorts of documents, including these:

• Last week, the Center for Public Integrity launched a series of articles on hidden hazards at oil refineries in the United States. Readers of Regulatory Flaws, Repeated Violations Put Oil Refinery Workers at Risk can review a dozen citations and court filings that the Center's journalists used in the reporting.
• Sunday, the New York Times published the first installment of an investigation into lax regulation of natural-gas drilling across the U.S., accompanied by a large cache of EPA and industry documents.
• The Seattle Times reported last week on evidence of financial abuse in Seattle public schools, based on documents released by state auditors. The documents detail over-billing, intimidation, and ethics violations that add up to $1.8 million in potentially fraudulent expenses.
• Statesman Journal reporter Tracy Loew used DocumentCloud to organize and annotate hundreds of documents that she accumulated in the course of her 16-month investigation of questionable spending in Willamette, Oregon's regional school district, WESD. Her reporting for WESD's Web of Deals turned up a decade's worth of red flags that had been raised -- and ignored -- in the agency's financial management, adding up to millions of misspent dollars. Statesman Journal used DocumentCloud to publish 300 contracts, court filings, and correspondence.

Remaking History

Documents in our catalog reach back into the past, as well. In 1970 Ruben Salazar was killed by police while covering an anti-war protest in east Los Angeles. A story rife with controversy, questions, and suspicions, his death became a rallying point in the Mexican American civil rights movement. Forty years later -- after refusing a public records request for documents that might shed some light on the circumstances of his death -- the Los Angeles County Sheriff's Department agreed to turn the files over to the Office of Independent Review.

While Los Angeles Times reporters waited for the report, they assembled their own folio of early clippings on Ruben Salazar. Readers can review FBI files obtained by the Times in 1999 and LAPD records on the department's repeated clashes with the journalist as well as a draft of the report prepared by the Office of Independent Review.

Join the Cloud

You can browse recently published documents by searching for "filter: published" or read up on other searches you might want to run. Here's hoping that the next year brings millions more pages, and more great document-driven reporting.

Spot.Us founder David Cohn has convened a virtual carnival: He's posing monthly questions that he'd like to see journalists take a stab at answering. The latest: how do we diversify the news ecosystem? He put it differently -- "Considering your unique circumstances, what steps can be taken to increase the number of news sources?" -- but I'm pretty sure the end goal is a greater diversity of information and expanded news ecosystem.

What can I do? Work to make document-based investigative reporting a little easier and a little more transparent. That's what DocumentCloud is all about.

One thing we can do is make it easier for reporters to access information. That would certainly go a long way towards making it possible for more New Yorkers to report on New York City policy and governance.

Another thing newsrooms can do, and here is where understanding news as an ecosystem, built of organisms that depend on one another comes in, is to spend less energy competing to get a story first and start acknowledging reporting that inspired a story.

No one newsroom can ever tell a complete picture, but a complete picture emerges from the work of many reporters looking at the same story, each with their own unique perspective. If some of those reporters went to journalism school or were mentored by a prize winning journalist while others are just calling it like they see it without even the benefit of a copy editor, more power to us all. (And if you imagine that the former never get a story outrageously wrong or that the latter are never spot-on, you haven't been paying attention.)

We're primed to think that competition is a good thing, and sometimes it is. Sometimes competition is exactly what drives us to do our best work. Sometimes, however, newsrooms get so caught up in the race to get to a story first that they forget to give props to the inspiring work happening around them.

Neighborhood Blogs are a Great Resource

One of my favorite neighborhood bloggers, who regularly reports on the shenanigans of local developers and local precinct community meetings -- stories no newsroom is watching nearly so closely, also keeps an unfortunate running tally of stories of hers that were picked up by the press without so much as a nod. All she really wants is a nod, but a news cycle stuck on getting the story first can't give her that.

There's nothing wrong with looking for leads neighborhood blogs. When you one, though, find a way to weave an acknowledgement into the story. It is common courtesy, and it leaves your friendly local bloggers free to be incensed by construction gaffes and intransigent municipal bureaucracies instead of ticked off at you.

Journalists ought to be able to share their reporting, pool technical skills and give one another the courtesy of due credit. All would go a long way toward increasing the vitality of the news ecosystem.

Gail Robinson's recent post on traffic in a post-loyal era got me thinking about measures of web traffic and, more broadly, how to measure the impact of non-profit journalism.

I certainly don't disagree with Gotham Gazette's decision to pass on providing Yahoo with free content. There's no good reason that Yahoo can't create a lively community without wholly reprinting Gotham Gazette's excellent original reporting free of charge.

There are probably good reasons that it would complicate Gotham Gazette's work to license stories to a commercial outlet like Yahoo Local, too: As a non-profit, the local policy publication regularly livens up stories by illustrating them with images licensed only for non-commercial use, or by independently licensing photos that aren't available under a Creative Commons license at all. Sorting out the images that can be re-licensed to a commercial entity like Yahoo isn't a trivial project, especially not for a small local publication.

It doesn't look like Gotham Gazette is alone in declining Yahoo's advances -- Yahoo Local's New York City page was recently dominated by pleas for piety from someone in Georgia:

And I definitely appreciate the impulse to own your traffic. One of the reasons DocumentCloud is thriving right now is that we've been very careful to ensure news organizations aren't handing traffic off to us. They own their traffic. They can keep track of their readership numbers, evaluate efforts to increase site visits, and slap as many ads and extra navigation elements on embedded documents as they want. Even so, they want more: Users and prospective users alike regularly ask for better metrics on the documents they're publishing.

Meaningful Metrics

Oakland Local, a project as commendable for its willingness to share insights as for its local coverage and community, has been quite open about the stats they look at as meaningful: Page views, unique visitors, average time on site and returning traffic. Returning visitors made up half their traffic when they spoke with Michele McLelland last spring. They also keep an eye on where their readers are coming from -- they're interested in how much of their audience is reading from Oakland.

When I was at Gotham Gazette, in addition to those basic web analytics, I kept a close watch on our comments -- their vibrancy struck me as a good measure of participation.

So what do you measure?

So I'm curious: Do you look for measures of your impact beyond the kind of numbers you show to advertisers? Share your thoughts in the comments below.

When we embarked on the DocumentCloud project, tools for altering documents were the furthest thing from our minds. After all, a responsible journalist doesn't tweak source documents!

But one of the first papers to embed material using DocumentCloud needed to do just that. The Chicago Tribune accompanied their coverage of a troubled foster home with a collection of letters and court orders. Though the documents offered an excellent illustration of the state child services agency's lax oversight and slipped follow-ups, they were predictably full of personal information about children in the foster care system, individual agency staff names and other personal and identifying details about private individuals that the Tribune opted to omit from their reporting. That decision, however, left the news apps team replacing the whole stack of letters multiple times before the package was finally ready to post.

A tool, right inside of DocumentCloud, for replacing, removing and re-ordering the pages of a document would have helped them a lot.

When the "PBS NewsHour" shared a century old hand-written Mark Twain essay, our OCR tools were not nearly up to the task of reading his handwriting. NewsHour transcribed the 10-page essay by hand and we worked with them to replace the text stored in DocumentCloud and displayed on the embedded letters.

By the time that Memphis' Commercial Appeal wanted to run a lengthy series of handwritten letters from Abdulhakim Mujahid Muhammad, a young Memphis-born man who opened fire on a military recruiting center in Little Rock last summer, we at DocumentCloud were busy supporting nearly 200 newsrooms -- offering to hide the text tab was the best we could do.

What NewsHour and Commercial Appeal really needed was a tool, right inside of DocumentCloud, with which to edit the text of each document.

And so, we've assembled what we think is a sweet suite of tools to let you re-order pages, insert new ones, delete old ones and edit the text that will appear in your embedded document. Check out our user guide to see how it all works. We welcome your bugs, feedback, rants, raves and, as ever, your documents.

Planning to spend the long weekend finalizing your Knight News Challenge application? It's too late for my favorite bit of advice ("don't wait until the last minute!"), but as someone who's been involved with three different winning projects, I like to fancy that I've got got some insight into what makes a good project.

A half dozen prospective applicants have sat down with me to workshop their News Challenge ideas, and I think I've helped them think through their projects to get them to a more viable place. The application process isn't hard, but you do need to give some sincere thought to your project or you're just wasting your time. Here's the advice I keep giving people:

Tips for the News Challenge

Focus on work you really want to do -- If you have a great idea but aren't really personally invested in making it happen, you're going to face a long, long slog if it gets funded. Three different people have complained to me that software developers put a ton of time and energy into developing Knight proposals that didn't wind up getting funded. That's always a let down, but it shouldn't be the end of the world. If you do make it past the first round, Knight is going to ask you a lot of hard questions and work with you to revise your proposal. If you don't get funded, you're left with a pretty solid and well thought out proposal that you can shop around if you really want to raise the money you need to get funded. That's a good thing!

Tell a solid story of engagement -- I'm not an expert on what is and isn't news, and I cut my teeth most recently at Gotham Gazette, which has pretty distinct standards for what qualifies. The most fascinating story won't find a home there if it doesn't have any apparent policy implications. I'm pretty sure that Knight doesn't look for policy implications alone, but if you can't tell me a solid story that takes me from your project to citizens (and non citizens!) and helps make them more engaged in decision making in their communities in some tangible way, something is missing from your project. That, or you're making me work far too hard to understand why this matters. So spell it out.

Make a realistic budget -- Grants awarded by the news challenge vary wildly in amount. Meaning: You should be honest with yourself about what it will cost to see your project through. A low ball request could leave you without enough money to finish what you started, and could be a sign to Knight's reviewers that you don't have a good understanding of what your project is going to take. A stratospheric budget isn't any more realistic.

Have a realistic outreach plan -- If you've got a great idea but no idea how to connect with the users that should be taking advantage of it, you've got a silo. Think this will be useful to a community? Go out and talk to people about what you're trying to do, how you think it will help and listen to what they say about how they want to use it. Not just what they think would be nice for other people to use, but what they want and will use.

Proofread! Proofread! Proofread! -- Your goal is to impress people, and to impress upon them that they should take a chance on your bright idea. Attention to grammatical details doesn't matter to everyone, but to some people a misplaced modifier is like nails on a chalkboard. Why risk alienating a reviewer?

Don't give up -- Knight's reviewers are going to look at zillions of proposals. If you're convinced that yours is a good idea but Knight turns it down, don't just quit. Keep looking for ways to make it happen, and keep listening to your community for insights that might make it a stronger project next time.

PS. I really have been involved with three Challenge winners. I wasn't around to help write the proposal, but I joined Gotham Gazette full time as director of technology after they won a 2007 news challenge grant to develop a series of games about public policy. Two years later, I helped develop Gotham Gazette's winning Councilpedia proposal before I joined the DocumentCloud team.

When we make lists of the kinds of source documents users can upload to DocumentCloud, they can get pretty long. DocumentCloud is court filings, hearing transcripts, testimony, legislation, lab reports, memos, meeting minutes, correspondence. I can say with absolute confidence that in all of our planning, "ballots" never once came up as the sort of document a news organization might want to annotate for readers. Our relentlessly creative users have shown us otherwise.

This summer, the Memphis Commercial Appeal rounded out its guide to August's primary elections with a sample ballot. Their digital content editor told us that many readers who'd missed the sample ballot in the print edition turned to the version online as primary day approached. Earlier this month, they added the general election ballot to that guide.

New York Ballots

WNYC, New York City's NPR affiliate, also published a few ballots this summer. In an effort to comply with a 2002 federal law that mandates significant updates to voting systems in each state, New York City introduced paper ballots for the 2010 primary election, replacing the city's famously arcane voting machines. One look at the new design and everyone was up in arms, proclaiming its absurdity, but WNYC actually invited a group of ballot design experts to review the city's new ballots. Their findings: the ballot was confusing.

Design for Democracy works to increase civic participation, in part through a ballot design project that aims to make voting easier and more accurate. WNYC used Design for Democracy's feedback to annotate a sample ballot on their blog, offering readers vital voting advice.

When the city released sample ballots for November's general election, a local think tank pointed out that the instructions erroneously advise voters to mark the oval above their candidate's name. In fact, the relevant ovals appear below candidate's names. WNYC highlighted the issue by embedding a sample ballot on their blog. Apparently the "oval above" language was mandated by state law. Don't believe me? See for yourself -- WNYC posted the legislation, with the relevant passage highlighted.

From now on, my laundry list of things DocumentCloud catalogs will most definitely include ballots.