Category: data

Categories
architecture Best practices data development Standards Web Standards

Understanding MARTI: A New Metadata Framework for AI

At its core, MARTI is a bridge. It harmonizes with existing metadata standards like the Content Authenticity Initiative, Anthropic’s Responsible Scaling Policy, and the W3C’s PROV. It anticipates the needs of future standards, laws and practices, such as those proposed by the Coalition for Networked Information (CNI), The EU Artificial Intelligence Act, and Making Data FAIR.—Carrie Bickner

As I study Carrie Bickner’s initial posts on the MARTI Framework she’s developing to manage AI metadata across various disciplines, a familiar feeling steals over me.

It’s similar to how I felt during the early days of The Web Standards Project (WaSP), when a handful of us took on the quarreling browser makers in what seemed a Quixotic attempt to bring consistency, predictability, usability, and accessibility to an already Balkanized web.

Fortunately, at that time, we had two aces up our sleeves: 1., the standards already existed, thanks to the W3C, and 2., the EU and Clinton Administration were suing Microsoft, which meant that the tech press was interested in hearing what we had to say—even if evangelizing web standards had little to do with accusations that Microsoft was abusing its monopoly power.

Once more with feeling: standards from the community

Years after The WaSP declared victory, and browser stagnation had begun to set in, I felt that same thrill vicariously when Eric Meyer, Tantek Çelik, and Matt Mullenweg invented XFN (XHTML Friends Network), inverting the standards creation pyramid so that great ideas were empowered to bubble up from small groups to the wider community, Open Source style, rather than always coming from the top (W3C) down.

I’ve no doubt that microformats were the spark that lit the HTML5 fuse, and we all remember how Steve Jobs used the new markup language to power the first iPhone, initiating the mobile era we now live in.

More about microformats history is available, and you can read Jeremy Keith’s HTML5 For Web Designers online for free—or buy the 2nd Edition, coauthored with Rachel Andrew, directly from Jeremy.

And now I feel those same stirrings, that same excitement about possibilities, as I study Carrie’s first posts about MARTI, an emerging object-oriented metadata framework that can be used to articulate rights-permissions, preservation metadata, provenance, relationships between objects, levels of AI involvement, and contextual information such as usage history and ethical considerations. 

Here’s why I’m excited (and you may be, too).

What do you wanna do tonight, MARTI?

For better or worse, our ideas create our reality. For better or worse, we have atomic power, the web, and social media. There’s no putting these genies back into their bottles. And there’s certainly no shutting down AI, however you may feel about it. Nor need we, as long as we have smart guardrails in place. 

I believe that MARTI—particularly as it promotes responsibility, transparency, and integrity in documenting AI’s role in content creation and curation—has the potential to be one of those guardrails.

Drafted by a career digital librarian, this provisional  metadata framework for human/generative AI output won’t stop bad actors from scraping content without permission. But if it is extended by our community and embraced by the companies and organizations building AI businesses, MARTI has the potential to bring rigor, logic, and connectedness to the field. In Carrie’s words:

The emergence of generative AI marks a transformative moment in human creativity, problem-solving, and knowledge-sharing. MARTI (Metadata for AI Responsibility, Transparency, and Integrity) is a provisional metadata framework designed to navigate this new landscape, offering a standardized yet adaptable approach to understanding, describing, and guiding the outputs of human-AI collaboration—and even those generated autonomously by AI.

At the heart of MARTI lies a robust object model—a modular structure that organizes metadata into reusable, interoperable components. This model ensures transparency, traceability, and ethical integrity, making it the cornerstone of the MARTI framework.

MARTI is not just an architecture for describing AI output, but it offers a way of structuring policy and a possible foundation for a new literacy. This is not about teaching every individual to code or engineer prompts. It’s about empowering humanity to collectively understand, describe, and guide everything we make with AI, ensuring accountability, transparency, and ethical integrity at every step.

MARTI is a framework for creating structured, standardized documentation that is attached to or embedded in AI-generated content. This documentation, or metadata, can be created by people collaborating with AI tools to produce content. Additionally, AI processes themselves can generate and embed metadata into their outputs, ensuring transparency, traceability, and accountability at every stage of content creation.

MARTI also offers a variety of potentially transformative business applications.

Disclaimer: the author is a friend of mine. But then again, so is every other thought leader mentioned in this article (with the exception of the late Steve Jobs, although our lives did touch when he fired me from a project—but that’s another story).

For more MARTI magic, check these posts:

And if you’ve a mind to do so, please pitch in!

Categories
A List Apart content strategy data Publications Publishing Stats

A List Apart: Just the Stats

Continuing with our “data, and what we can learn from it” theme, here are A List Apart‘s four most popular individual pages this week (excluding the home page, with 178,270 page views). Pay particular attention to the publication dates:

Article Page Views This Week
Flash and Standards: The Cold War of the Web by Daniel Mall, MARCH 9, 2010 41,035
Drop-Down Menus, Horizontal Style by Nick Rigby, JUNE 29, 2004 40,683
CSS Design: Taming Lists by Mark Newhouse, SEPTEMBER 27, 2002 37,867
[Articles Index Page] 34,630

What do these stats tell us?

For one thing, they tell us that for every reader who viewed A List Apart as a topical periodical publication—that is, for each person who read one of its latest articles—there were two readers who used the magazine as a source of evergreen web design and development content.

Put another way, for every person who uses A List Apart like a magazine or blog, there are two who use it as an encyclopedia of best practices in coding and design.

We’re looking at only seven days worth of statistics, here, but the pattern is consistent from week to week. What it tells the editors is that we’re not in the quick-hit eyeballs and ad sales business (but we knew that), we’re in the professional education business. It reminds us, if we needed reminding, that the mission stated on our Contribute page is still true:

We want to change the way our readers work, whether that means introducing a revolutionary CSS technique with dozens of potential applications, challenging the design community to ditch bad practices, or refuting common wisdom about, say, screen readers.

As we go about soliciting and reviewing potential ALA content, we must keep uppermost in our minds how people use our site, because its value lies in the hardiness of our best articles. Web professionals trust us to have the information they need to do their jobs better and deliver the best possible experience to their clients’ customers (and thus the best value to their clients). These stats reinforce what we already know, and help us stay true to our core purpose and values.

This emphasis on the evergreen over the topical also helps us as we strategize means of deepening our relationship with the web design community and add or change features to do so.

I’m barely scratching the surface of what the data tells us, but the little I have teased out so far is already very useful—and that’s the point. The more you look, the more you can learn. What’s in your logs?

Categories
Community content content strategy data Design Stats Zeldman zeldman.com

Love Me Long Time

Those who say web users don’t spend time reading web pages haven’t met readers like you folks. According to Google Analytics, zeldman.com fans spent five minutes, fifty-five seconds reading the relatively short post, “My Love/Hate Affair With Typekit.” If Jakob Nielsen is right, and readers take in no more than 20% of the words on a page, y’all took a hella long time to read 190 words.

But generalized findings like Jakob’s are merely one data point in a universe of possibilities. Every site is a special snowflake, with stats and usage patterns all its own. Faced with an unfamiliar shopping site, we may indeed give it little more than a cursory scan before closing the window and returning to Google to fine-tune the search that led us there. But when we visit a familiar site to read, then read we do—as anyone with a good blog and a decent set of analytics tools can tell you.

Here are a few recent average times readers spent poring over various zeldman.com posts:

Post Title Average Time Spent
My Love/Hate Affair With Typekit 5:55
Crowdsourcing Dickens 3:36
20 Signs You Don’t Want That Web Design Project 7:52
Ed Bott’s Lament 4:22
Gowalla My Dreams 4:41
IE9 Preview 4:37

Morals of the story:

  1. Don’t use Peter’s stats to paint Paul.
  2. If you want people to spend time reading your site, give them better content.
Categories
A List Apart data Publications Publishing

Future of Online News

Many website designers run their own niche blog, and if the content is unique enough, a designer might be able to sell subscriptions to it. The content has to be very high quality, though, and few design blogs meet that standard.

A List Apart is one that does, and it could potentially turn a profit selling subscriptions. But the subscription route is a risky move because it alienates many users and shrinks ad revenue substantially.

Jeffrey Zeldman, publisher, founder and executive creative director of A List Apart, gives two reasons why A List Apart does not put its content behind a paywall:

  • It’s against our belief in free online content.
  • It wouldn’t work unless our competitors also put their content behind a paywall. We appeal to a discerning base of web designers, but if we went behind a paywall, it would be as if we had stopped publishing. Our readers would turn elsewhere.”

More at What Is the Future of Online News? | Webdesigner Depot.

Categories
Accessibility Advocacy Blogs and Blogging business Community content strategy data Formats glamorous HTML Ideas industry Publications Publishing Respect State of the Web The Essentials The Profession W3C work writing

Posthumous Hosting and Digital Culture

THE DEATHS of Leslie Harpold and Brad Graham, in addition to being tragic and horrible and sad, have highlighted the questionable long-term viability of blogs, personal sites, and web magazines as legitimate artistic and literary expressions. (Read this, by Rogers Cadenhead.)

Cool URIs don’t change, they just fade away. When you die, nobody pays your hosting company, and your work disappears. Like that.

Now, not every blog post or “Top 10 Ways to Make Money on the Internet” piece deserves to live forever. But there’s gold among the dross, and there are web publications that we would do well to preserve for historical purposes. We are not clairvoyants, so we cannot say which fledgling, presently little-read web publications will matter to future historians. Thus logic and the cultural imperative urge us to preserve them all. But how?

The death of the good in the jaws of time is not limited to internet publications, of course. Film decays, books (even really good ones) constantly go out of print, digital formats perish. Recorded music that does not immediately find an audience disappears from the earth.

Digital subscriptions were supposed to replace microfilm, but American libraries, which knew we were racing toward recession years before the actual global crisis came, stopped being able to pay for digital newspaper and magazine descriptions nearly a decade ago. Many also (even fancy, famous ones) can no longer collect—or can only collect in a limited fashion. Historians and scholars have access to every issue of every newspaper and journal written during the civil rights struggle of the 1960s, but can access only a comparative handful of papers covering the election of Barack Obama.

Thanks to budget shortfalls and format wars, our traditional media, literature, and arts are perishing faster than ever before. Nothing conceived by the human mind, except Heaven and nuclear winter, is eternal.

Still, when it comes to instant disposability, web stuff is in a category all its own.

Unlike with other digital expressions, format is not the problem: HTML, CSS, and backward-compatible web browsers will be with us forever. The problem is, authors pay for their own hosting.

(There are other problems: the total creative output of someone I follow is likely distributed across multiple social networks as well as a personal site and Twitter feed. How to connect those dots when the person has passed on? But let’s leave that to the side for the moment.)

A suggestion for a business. Sooner or later, some hosting company is going to figure out that it can provide a service and make a killing (as it were) by offering ten-, twenty-, and hundred-year packets of posthumous hosting.

A hundred years is not eternity, but you are not Shakespeare, and it’s a start.

Categories
data industry Information architecture State of the Web

Men like it fast, women like it good

In a recent usability survey, researchers from Southern Illinois University found that after ease of use, men prefer fast download speed to easy navigation. Women prefer ease of use, easy navigation, and accessibility. The researchers hypothesize that these different usability criteria are due to differences in how men and women use the web.

Details at “Usability Study: Men Need Speed – web usability criteria show gender differences.”