Using the HTML Meta Pro, Pro-Auth, Pro-Auth-Field, and Pro-Auth-Fragment Tags
Abstract: New meta tags will flag authoritative content for potentially higher ranking in search engine results. Two are “pro-auth” and “pro”. Two more are “pro-auth-field” and “pro-auth-fragment”.
Failed Searches Reveal a Problem
I searched in Google. The first result that came from a domain I trusted was the 17th result. It was on page 2 of the results. That’s too far down. Too many results I didn’t trust, sixteen in a row, were higher. It was too easy to get a wrong answer.
This is a problem especially for subjects that are critical for people’s lives and opportunities, like health and law. Plenty of websites look dubious or are plainly wrong, on those subjects and probably on all major subjects, and those sites often top the search results. It’s also a problem when the high-quality content is only on certain pages of a website and the rest of it is virtual garbage. An example of this could be a business that sells, say, many tools that are not very good, filling many pages with them, but has a couple of excellent discussions on how to build things.
Google tries to address this, in certain fields, by looking for signs of credibility, but it’s not always successful. I don’t know if any other search engine even tries.
One solution could be to seek out relationships of authority. A professional organization might believe its members and allies. Scholars may respect the high standards held by many professors, students, and graduates. Journalists with specialized beats can rely on some people who are openly sources of information and can rely on other journalists. A committee that awards prizes to the best in a field might have earned respect for its decisions. A peer-reviewing publisher would likely believe its authors. A highly-regarded art critic can commend artists and musicians for the public to take in. A fan club can point to well-informed media. A government licensing agency recognizes licensees as typically having authority held by no one else in its jurisdiction. A mafia can list its skillful underworld affiliates who host pages of know-how (right now, maybe a police detective is saying, “oh, please do”). An enterprise can list its expert employees and contractors, like a hospital linking to its doctors’ Web pages. A car company can link to independent mechanics who have pages of good repair advice; a home supply store can link to informative plumbers and electricians.
No limit exists, essentially, on who can have the authority to judge who else can be trusted to be reasonably informative on a subject. And, likewise, that authority need have no limit on whom they can decide is an expert in their shared field. Trust can be offered downwardly, upwardly, or laterally, e.g., to colleagues. Affiliation by other means is not required. So, for example, a medical organization can decide that a mass-media newspaper headlining Hollywood gossip nonetheless had a great article on cancer, even though there’s no affiliation between the newspaper or its reporter and the medical organization.
This is not meant to link to all of your employees, contractors, creditors, investors, and customers or everyone you ever heard of or everyone in Google. This is for experts who post some of their expertise on the Web.
Manual research uncovers some of these relationships today. If you know nothing about widgets, you can still research the field, try to discover one or two names of bodies or people who generally know a lot about widgets and won’t mislead you (provided they’re willing to talk with you), and, in many cases, find out whom those one or two trust in turn. But a problem with this method is that it is slow and demands a lot of labor.
Relating Two Sites Together
A faster way would be to let some of these relationships be public in a way that a search engine can discover before you need them. The search engine can then evaluate them before you decide whether to rely on them. Even if the search engine has its human staff do the evaluations, that can be faster, done sooner, and more efficient than waiting for millions of users each to try it from scratch.
Directories of the Web can turn to this feature, too. Anyone can, especially if they visit websites, read the tags there, and analyze multiple websites for relationships.
An offline authority or offline expert is beyond the reach of this system. Either one would have to be at least referenced online, on a website.
The Authority on Everyone Else’s Expertise: The Pro-Auth Tag
First, we need an authority. This could be a university department; but it could also be an amateur, a hobbyist, or a fan who knows, or says they know, a lot, because this system has a way to judge that claim of expertise reliably.
Anyone declaring itself or themself to be an authority on any subject only requires the page author to add a meta tag with the “pro-auth” name into the source code of one page. Whether a search engine agrees is entirely up to the search engine. Maybe the subject is not important enough to require an authority. Maybe the subject is important but this organization or person is not the one to trust in that role. This is not a pass/fail test; any degree of authority can be adjudged, and that degree might be zero, in which case the tag would carry no weight at that search engine. Another search engine might decide differently.
No coordination or agreement is needed between two pro-auth websites on the same subject. Likewise for two pro websites, or two pro-auth-field websites, or two pro-auth-fragment websites. (You may need to coordinate between pro-auth and pro, but that’s different from lateral coordination for the same tags.) If someone on the other side of a mountain is an expert on the best restaurants, but so are you, you can both use the same tags for the same expertise. It doesn’t matter whether you’ve ever heard of each other or even seen each other’s work. However, whether a search engine’s managers believe the other person at all, or believe you at all, or to what degree, is up to that management. What you can do is create your website’s content to show that yours is the more believable website and hope someone running a search engine believes you. (And, if you want to coordinate with the character on the other side of the mountain, the two of you can go ahead.)
Those judgments would likely be based on knowledge about users, knowledge of society and of the subject, what’s known about the party behind the website, what’s on the whole website and in other publications and oral information, justified reputation, and other evidence. This might be decided via a combination of an algorithm and human judgment. One search engine’s decision need not matter to any other search engine. A search engine’s judgment one day can be changed any time and rarely or frequently, at the engine’s whim. A specialized search engine may give more or less weight to an authority than a generalized search engine might.
Say the field of expertise is banking and a nation’s central bank lists its member banks. The search engine might give the central bank a lot of authority with respect to banking and give the member banks’ websites higher search result positions. But if the subject is flying saucers from Pluto, the search engine might give it no weight and then ignore the relationships. Every website about flying saucers from Pluto would be on its own, in the race for credibility, and could still appear in search engine results. It just would not have a relationship of authority accepted by the search engine. Anyone wanting to know if a saucer landing in their back yard came from Pluto could still visit the websites that’ll say one way or the other.
No negative evaluations are provided for with this tag. A doctor cannot make a list of quacks that this system would recognize so search engines would downgrade their search result positions. But the doctor can link in a positive way with this system to a list of quacks. And the doctor can simply publish a list of quacks without using this system for markup. Patients who use the Web can then try to select their health care providers. (Making a list like that may incur legal liabilities, which remain the responsibility of the one publilshing the list.)
Completing the tag requires the content attribute. That value would generally be a partial URL for the page where the tag appears, a fragment of that page, any portion of the website that includes that page, or the whole website that includes that page. The protocol and the leading punctuation would be omitted from the URL, since it is not meant to be a link per se, but an identifier. The base element would not be used as part of the partial URL. (An HTML link element is not useful here, even with a partial URL.)
About the Pro-Auth Authority and Why to Respect Its Pre-Eminence
For the qualifications to be an authority, what the “pro-auth” authority writes on that page is entirely up to the creator and the page author. It should be humanly parsable and need not be machine-parsable. The content about the authority might discuss what kind of expertise they have, how long they’ve been around, what recognition they’ve received themselves, how they’re distinguished from other entities, and references to literature substantiating (or perhaps criticizing) these claims. In other words, it should give information search engine humans can use to judge how much weight to give this authority. It can also serve other purposes. It could say that the authority may not have read the websites it lists and may not be aware of recent changes to any website. It could include disclaimers of legal liability for the purported experts and their purported expertise (lawyers are good at writing disclaimers).
If Only a Few People Should See the Pro-Auth List of Experts
If a “pro-auth” website is to hide a list of experts, such as behind a paywall, a search engine will need access and cannot be expected to enter passwords. A website creator may want to consider making the information public or, if some privacy is needed, relying on obscurity of a URL where the data is kept, even though that often is only a little bit private, and won’t stop many people from finding the information for free.
Compiling a Pro-Auth List of Experts: It’s Not Too Soon to Start
The “pro-auth” authority should immediately start listing whomever it decides has some of that authority and deserves to be trusted in the field, i.e., experts. That list, including links, should be on a page with the “pro-auth” tag. If it’s not, such as if it’s across multiple pages or needs a lengthy explanation, a page with the “pro-auth” tag should discuss the authority and refer the reader to the list or lists. The “pro-auth” tag can be on any number of pages, so, if a list is on another page, that page, too, should carry the “pro-auth” tag (along with a “pro-auth-fragment” tag, discussed below). If an explanation is on a page and the list is on three pages, the “pro-auth” tag should be on all four pages, but probably on no other pages. To help a search engine separate out information about the authority’s qualifications from the lists of links to experts, the “pro-auth” tag should also have the keyword “why” or “list”.
But Google, for example, does not like pages under construction, so, if a list cannot be compiled yet, adding the “pro-auth” tag probably should be delayed, or it should be removed, until at least a minimally adequate list is ready. Putting the tag in for an explanation without a list is generally not useful yet, so that tag can be deferred until a list is being posted. The list can be as short as one expert, as long as you have at least one.
Specialties of Authority: The Pro-Auth-Field Tag
The authority identified with the “pro-auth” tag should state the field of its authority, and may state more than one. It can do this by adding a meta tag with the “pro-auth-field” name for each field. The field would be identified with the content attribute’s value, one field per “pro-auth-field” tag (there could be many “pro-auth-field” tags). The content value would say whatever field, such as “contract law” or “Smith & Co., Inc., history”, that the page author or creator wants. There is no rigid list of fields, since tag writers should be as precise and accurate as necessary (which is why comma-separating or space-separating fields in one tag is not practical but writing multiple tags is allowed). With this tag, a search engine could handle the subtle problem of classifying a page as being recognized for expertise on mainstream medicine or alternative medicine, sometimes a difficult distinction between fields, by relying on what the “pro-auth-field” tag says in naming a field that suitably describes the “pro-auth” list that lists the page. Whether to believe that the authority really is an authority in the claimed field would be up to a search engine or its people to decide. The search engine could also decide to give heavy weight to one claim and little or no weight to another claim by the same website, thus agreeing on authority in one field but not in another, even though the same party is claiming both.
Separating Lists From Chaff and From Each Other: Fragments and The Pro-Auth-Fragment Tag
We don’t want to confuse links in the authority’s list or lists with other links that don’t share that authority. To prevent that error, the list itself should have HTML markup to distinguish it from other links. Specifically, each of these lists should have an HTML fragment identifier that nothing else on the page has, and this fragment identifier should be stable over time. What fragment identifier to give should be left up to the page author, because it’s impracticable to standardize it across the whole Web.
But a search engine would still need to know which fragment identifier is written for this purpose. So, to help the search engine, that fragment identifier should be copied into a meta tag in the source code. That meta tag would be named “pro-auth-fragment”. The content value would be just the fragment identifier, with a leading hash mark. More than one list can be given different fragment identifiers and they can all have matching meta “pro-auth-fragment” tags.
When a search engine discovers the page with a “pro-auth-fragment” tag, it can start giving links in that list the weight it assigned for the “pro-auth” tag. If the page is without links, the search engine’s bot probably should revisit the page in a few months, to give the main authority time to populate it.
(The leading hash mark is required, in order to reserve a space-separated string without a hash mark for a future purpose.)
Experts, Pretended Experts, and the Worst: The Pro Tag
At the other end of the relationship is a purported expert. That expert’s page author could add a meta tag with the “pro” name and associate the tag with the “pro-auth” page that recognizes this expert’s expertise. The way to associate it would be to copy the partial URL that’s in the authority’s “pro-auth” tag’s content value, and copy it into the “pro” tag’s content value. A search engine could notice the widespread use of the same partial URL in a “pro-auth” tag and many “pro” tags and build the authority relationships accordingly. It could even notice a simple 1:1 relationship.
Expertise may reside wherever it’s found, sometimes in surprising places. An authority can link to entire wesites or folders within websites, but it can also link to single pages that it considers good within websites that it considers bad. Even a piece of a page could be good enough. The authority could link to a portal, if it believes that what the portal itself links to is of high enough quality.
One expert might be supported by several authorities. Perhaps someone has a website with some history about Bentley cars (okay, automobiles). The manufacturer might be one authority and a fan club (okay, motorists’ association) might be another authority. Search engines can judge each authority relationship separately, and if for one the authority is weighty then that relationship could bring the website up in the search results. If for the other authority there’s no weight, the website could still get the boost it earned from the one good authority, because it need not be reduced or penalized for claiming the other.
It could also reject a relationship. If, for example, someone claims to be an authority on annual abductions of everyone in Miami Beach to Mars and says NASA is an expert on this and NASA says no, not us, NASA can deny the relationship to the so-called authority. It might want to if it’s getting an expensive amount of traffic from that claim through a link. NASA can do this by adding the meta “pro” tag pointing to the unwanted authority (by copying the partial URL) and adding the space-separated keyword or token “false”. This also applies to subdivisions within what the partial URL refers to; for example, if the partial URL is for a whole website, this applies to every directory, page, query string, and fragment identifier within the site. Google and others can discover this “false” keyword and not sully NASA’s reputation. The link would still exist; but in the absence of the authority relationship the search result positioning might drop lower. Political parties and campaigns, if finding damaging links from opponents, might find this feature useful.
If the page author of a page that was honored with an authority relationship decides to rewrite and downgrade the content, and doesn’t believe it should be trusted as much for its expertise, it could break the relationship by removing the relevant “pro” tag. The next time the search engine indexes that page, the break in the relationship can be noticed.
The “pro” tag being omitted or having the “false” keyword has the same meaning.
The “pro” tag having the “true” keyword means the same as that tag with the partial URL. The “true” keyword would be superfluous, so it would be pointless to use it (except perhaps for the sake of style).
Spamming search engines with an implied overclaim of expertise in everything would be counterproductive. While young children might make that claim and we’d often praise them for being so smart, an adult doing this is likely to be a moron or a liar and we don’t want to encourage that. This can be prevented. The “pro” tag having the content set as if to accept recognition from any “pro-auth” page would be void. A null value or a wildcard by itself is not allowed. Even a wildcard for part of a domain is not allowed. A partial URL, without a wildcard, is required. Multiple “pro” tags, each with a different partial URL, are permitted. (A question mark introducing a query string is still allowed in a partial URL, but not a question mark as a wildcard.)
Keeping the relationship secret from the public, while letting search engines know, is not possible in this system. Most people’s browsers allow viewing the source code of the current page, thus exposing the relationship.
Private use is possible, but not the norm. It is mainly for a single-institution terminal, an intranet not accessible from the Internet, or a computer that can be accessed using a loopback IP address (such a computer is sometimes named “localhost” or “localhost.*”). It would be better suited for where a separate search engine is installed, such as within a single company or on a single website.
In the End
It worries me a bit when someone tells me they got some important health information, I ask them where they found it, and they say “Google”. I like Google. I use it a lot. I use it for health information, too. But, substantively speaking, Google is not the source. The website that Google pointed to is realistically the source. For those of us who are concerned enough to try to limit what sources we’re using to those we’ll trust, an authority relationship is often helpful, sometimes even critical. These elements will go a long way toward making it easy to establish that relationship of authority.