gemini://the.web-is.fail/anti-user/
Anti-User Practices that SmolNet Solves
Technologically Impossible (or Nearly)
The protocols used in SmolNet are designed _not_ to support the following site behaviors. It may be that loopholes will be found (or wedged in) to make them possible, but by design they are not supported:
(NOTE: the impossibility of some of these behaviors, depends on what protocol you are using. The Gemini protocol is explicitly aimed at preventing most of them, whereas Gopher was only intentional about avoiding some of them)
No cookies
There is no way for a server to send you any kind of metadata other than the actual page content itself, and its content-type; nor for your client (browser) to hold onto anything from the server to be sent back to it (apart from URLs).
No automatic content-fetching
When the user requests a page, it is only that direct page, and nothing else, that is fetched and displayed. Not scripts, CSS, nor even embedded images (which, sadly, have historically been used to track you across sites). One page request results in (at most) one fetch.
In the case of the Gemtext format (Gemini and Spartan), embedded content is impossible, or at least quite impractical, as there is no real means of specifying it.
In the case of Gopher (and its index format), it would technically be possible for a client to automatically load any linked images (as those links are distinguished from other types), but possible or not, Gopher clients typically don't do it, generally speaking.
Even though images aren’t loaded automatically, you are still welcome to click a link to separately view any linked content you are interested in. Some Gemini clients (such as Lagrange) will display a clicked image as embedded in the page (since the fetched content can be determined to have been an image, _after_ having fetched it).
No cross-site tracking
As a consequence of the above, it is not generally possible for anyone outside of a site to track your path through it.
Note that a server can still track your navigation through it's own site:
- A server can track your navigation through its own site(s) by your IP address. If you use a proxy, anonymized NAT (network-address translation), or an anonymizing network such as the TOR (The Onion Ring) network, the server cannot track you, because the client cannot send any identifying metadata.
- A server could also track you by injecting tokens into the URLs that it provides to you.
- If you are using TLS-certificate identities, the server can obviously track any content you fetch under that identity. Being tracked is, of course, the actual *point* of using an identity: so that the server can verify that you are the same person every time you use it.
Regardless, no server can track what you do on some *other* server—at least, not without that server’s explicit cooperation.
No “fingerprinting”
Targeted advertising on the web has never been more sophisticated, aggregating tiny, seemingly-insignificant points of data to form an identiy profile. Using things like what browser and OS you’re using, small differences in how your browser behaves, little differences in how *you* behave, the set of fonts you have installed, etc, modern sites can still track you even if you’ve blocked cookies!
None of this information is available to servers via Gopher, Gemini, or other SmolNet protocols—no fingerprinting on the Smolnet!
No doom-scrolling
One URL visit = one page-fetch; no further content can be fetched without the user clicking a link.
It _is_ possible (for some clients that support it) to stream content without a defined end, which is somewhat similar to “scrolled content”... except that a website that provides the “doom-scrolling” experience needs to detect when you are at or near the end of the content it has provided until now, so that it knows to load more content. The SmolNet protocols are not capable of informing the server where you currently are on the page, and so doom-scrolling is impossible.
No preventing “Save As...” or “View Source”
On the web, some sites will use scripts in an attempt to prevent right-clicking, and thereby prevent you from inspecting the page’s raw source, or from downloading any embedded images to your local disk (where you might—gasp!—continue to view it without having to go through their platform!)
On the SmolNet, there’s no scripting, and there are no (automatically) embedded images anyway, so there’s no way to prevent downloads or views (and... not much content to prevent download of anyway).
No content hiding
No scripting, so no article pages that teasingly let you see the content briefly, and then hide it by locking the scrollbar and popping up a subscription box you can’t dismiss (without subscribing).
In theory, a server could just serve an excerpt of the content, and then make you pay to see the rest... not much you can do about that technologically, though. At least they can’t play “now you see it, now you don’t!”
Know when you’re leaving safety!
Not all clients, but at least several, will clearly mark any non-SmolNet web links, by coloring them differently, or by using different symbols for various link types, so that you know if you are leaving the guarantees of the SmolNet for the uncertainties and complexity of the Web. (And, of course, there’s no means for a SmolNet page to interfere with that.)
Not Impossible, But Absent Nonetheless
There is nothing in the way of technology preventing the following kinds of behaviors, except the goodwill and principles of the community itself. But, they are nonetheless absent (as of the time of this writing, to the knowledge of the author). So long as this remains true, they constitute additional blessings that can’t be depended upon on the World-Wide Web, but that you may be confident that you will not encounter on the SmolNet (...so far, and for now).
Intrusive advertising (or, advertising)
It would be possible to inject text-only advertising into page content. It would be particularly gross If such ads were to masquerade as legitimate content. It might be hard to avoid at least simple, clearly-marked text ads of any kind at all in the future, if the SmolNet achieves even a hundredth of the attention that the web has. But it doesn’t, and commercial ads are either non-existent, or nearly so.
Also, a clear disadvantage to text-only advertising, from the advertiser's point of view, is that it becomes difficult to measure or confirm that the sites doing the advertising, are truly being served up as often as the site's management is claiming. Nor can you check such things as “mouse hovers”, nor how long the page's scrollbar is lingering where the ad is within view.
Platform entrapment
Out on the web, it has become commonplace to build large, centralized platforms with attractive features and liberating, users-first policies, unencumbered by excessive advertising... only to later snap the trap-jaws shut when it’s too late, and everybody’s already there, and completely dependent on its conveniences. Then, they lock down the platform and tighten restrictions, become a firehose of advertisements, institute anti-user polices designed to keep you only on that platform, and minimize the opportunities to visit others, or sometimes even prohibit linking to them.
Even the content that you’re there for, will become less accessible, as they place obstacles between you, the user, and the content you’re actually there for in the first place. After all, the more time you spend trying to get to the content, the more opportunities they will have to advertise at you, or otherwise leverage your attention.
Nothing about the SmolNet prevents this from happening, apart from the collective intolerance of the community. But it hasn’t happened yet.
SmolNet can’t prevent it happening. But a dogged unwillingness to participate in large, centralized platforms, and a strong aversion to decentralized technologies, could!
Deliberately inefficient feed algorithms or searches
There is, of course, nothing preventing a server from making itself less reliable/less helpful to the user who wishes to find or discover something. However, I think there are two things that help limit the motivation to do so.
For one, current search and feed technology used on the SmolNet is exceedingly basic. As far as I can see, if Google were to make the choice to make their search functionality deliberately worse (as some have reported they've done), it would only be because they are head, shoulders—and waist!—above their competitors in being efficient, and have ample room to degrade search performance in the interest of ad performance while still remaining far better than their competitors.
For another... if the motivation to keep you searching or navigating longer is that they can serve you more ads, then the opportunity for that is fairly limited. As mentioned, it's not technically impossible to serve ads on the SmolNet, but it _is_ impossible to serve unasked-for images... which only leaves text insertions. In my view, there is a much lower limit to user tolerance for ads being inserted into text content, and if you were to combine that with a search site that has you viewing more pages than you'd otherwise need to, the tolerance would be lower still.
But that's an opinion, and it may well be wrong. Perhaps all the necessary motivations and opportunity exist, and we're saved merely because the current SmolNet culture is, momentarily, both inclined not to create such devils, nor tolerant of them were they to be introduced. The result, for now at least, is the same: you won't find it here.
--
Back to the start page
Copyright (C) 2026 Kado <---@---.--->
If you email me about this site, please include "WEBFAIL" in the subject.
This site (the.web-is.fail) is licensed under the CC BY-NC-SA 4.0.