archives browsing by category


Holovaty: We're building databases for the future, not the present

Thursday, September 7th, 2006

As you may have heard, a fellow named Adrian Holovaty has a Big Idea, and it’s a really good one: newspaper information needs to be updated for the digital age by storing it not only in the hundred-year-old “story” format, but in little database chunks. What’s the business model? He’s quick to say he doesn’t know, but his answer this week to one concern should be comforting to data compilers with small audiences.

Here’s Adrian’s line: if you’ve lifted a few words out of your story and flagged them in a way that a computer can recognize — if a computer can indentify your story’s “who” and “where” — then you’re setting yourself up to someday ask a computer to map all those “who”s against, say, a database of political donors, or real estate purchasers, or sources. It’s the difference between Fisher-Price and Lego.

His most visible work at the Post has been stuff like the wonderfully nichey political ads database. But in a long blog post this week, he reminds us that mere newsy databases aren’t the endgame — the greater purpose isn’t serving today’s reader, but laying the ground for future remixes of the data.

If you store everything on your Web site as a news article, the Web site is not necessarily hard to use. Rather, it’s a problem of lost opportunity. … That Web site cannot do the cool things that readers are beginning to expect.

That’s a bit of encouragement for small papers considering similar projects in the face of minimal pageviews. (Squint. You see that? Wagging in the distance? It’s the long tail!)

It’s also some motivation to keep that data well-scrubbed: more’s at stake here than Saturday’s paper.

A subscription model that won't compete with print: the blindspot

Saturday, June 17th, 2006

Tear down the wall? At the Times, too early to say. But in the next few years, small papers should build their subscription strategy around this question: what on the Web is a substitute for print, and what’s not?

A model I like, but have never seen, is actually the inverse of the most common one. Instead of a permanent archive wall, it’s an ever-advancing blindspot.

For the next ten-to-15 years or so — until computers become almost as portable/cheap/comfortable as newspapers, that is — small newspapers should prioritize new editorial Web features with the following checklist:

1) Can it be done with information we already collect?
2) If not, can it be done with information whose collection is easily automated? (either through user contribution or computer algorithm)
3) Can it be presented in a way that is only possible or convenient online, so as to avoid substituting for the print product?

From this angle, charging for archives looks like the dumbest possible formula. We’ve all got colossal electronic archives. All we need to make them useful is a good search feature. And here’s the thing: archives don’t substitute for print at all. What subscriber saves two-week old newspapers for use as reference material? Online archives only add value. A free, well-ordered archive for a local newspaper would take it a long way toward its eventual goal: becoming the primary information site for its community.

Yesterday’s news is different. In most cities, you can get yesterday for 50 cents in the newspaper, or on the Web for free. Print and Web become substitute products — and get moreso with every redesign.

Okay, what about today’s news? I lean toward the Spokane model — breaking news and comment should be free. They’re dynamic. They can’t be done in print. They’re dealing with radio and TV competitors.

You can see by now what this all means: the sensible place for a subscription requirement is content from, say, the last three days. Farther back than that, it should all be free again.

I’ve never seen it done. I’d love to hear why not.

(Also: Yes, yes, I know, I should be preparing and posting my own archive of three-quarters-written entries instead of making a new one. Sorry, chum.)

Reinventing news for a search-based world

Sunday, February 26th, 2006

On OJR, Robert Niles has a fascinating suggestion for completely rethinking online news: replace inverted-pyramid newspaper articles with staff-written wikis. (For which he suggests the delightful nickname “stikis.” You heard it here first, dear reader.) Why? To attract search-engine traffic.

I’m suggesting that — instead of distinct daily takes — news stories could be covered with encyclopedia-style articles that staffers would update with new information whenever available. How many more inbound links would such an approach get?

Inbound links, of course, being the current currency of the search-driven Web.

I see two obvious problems with the suggestion:

1) the writing required would be so different from that used in the print product that editorial resources would be taxed, and
2) the lede would be perpetually buried; that is, readers would have trouble figuring out what parts of the news are new.

Though the second problem might be avoided with constant rewriting and clever formatting, my hunch is that self-contained news stories will remain the dominant delivery device for news. They’re simply easier to pluck relevant details out of. This isn’t to say that local papers couldn’t launch parallel stiki or wiki services for their coverage area. But I doubt extensive stikis are likely to be worth their while.

Using tags in local newspaper archives

Sunday, February 26th, 2006

How can small papers get the most out of their archives? Folksonomy is sexy, but most Web audiences are too small for it to work well. However, old-fashioned, top-down tags can be a useful tool for the many local papers that, like many Web newcomers, still sort their Web content with print principles in mind.

(Yes, has added links to all its stories. Yes, this is a terrific promotional move for a national paper. No, it’s doesn’t add as much to local papers, because they aren’t looking for national readers. But yes, local papers ought to consider offering it merely as a service to readers.)

But here’s something more important that small papers could easily do with minimal staff effort: make a list of a couple dozen possible “tags” for stories. When any story is sent to the Web, a staffer can glance over the list of tags and check any that apply.

A newspaper in North Carolina might write frequently about the pork industry. Does the print edition have a section entirely for the pork industry? Of course not. And as a result, the paper’s Web site doesn’t, either. But it should! A small set of staff-written tags are easy ways to build topical archives — whose index pages (available in RSS, of course) can then be used to cluster non-news content for the target audience, such as topical blogs, off-site links, and of course targeted ads.

What’s more, a tag system, unlike a print-style tree categorization, lets stories fall into multiple categories. And it can be used to easily locate similar articles, which have similar combinations of tags.

Readers should be able to navigate newspaper Web sites in various ways: search, vertical brosing (feature section->movies page->Brokeback Mountain review), and horizontal browsing (Brokeback Mountain review->Oscar preview). A staff-maintained tag system is one way to facilitate that.