What the browser is doing

An ordinary web address is turned into a network location by asking a directory. An address of this kind is not. The 56 characters are the identity, and reaching the service means finding a path to whatever holds the matching key rather than looking up where it lives.

That difference explains most of the behaviour people find strange. There is no directory to be slow, no cached lookup to be stale, and no helpful redirect when something is not found. There is either a path or there is not.

The browser gives you a progress indication, but it is a rough one, and it does not distinguish between the several very different things that can be taking time. Building a path takes time. Finding the service takes time. Waiting for the service to answer takes time. All three look identical from the outside.

This is the root of most confusion on this board. When a load is slow, the screen offers no information about which stage is slow, and people fill the gap with a story. Usually the story is about the destination being down, and usually it is unfalsifiable from where they are sitting.

• It will not correct a mistyped address. There is nothing to correct against.
• It will not suggest a similar address, because similarity is meaningless here.
• It will not tell you which of the three published addresses you are on unless you look at the bar yourself.
• It will not warn you that an address is not the one you used last time. As far as it is concerned each is simply an address.

Whichever nexus market onion address you used, look at the address bar after the page loads rather than before. It is the only moment where the string in front of you is the string that actually produced what you are looking at, and it costs a second.

Compare the part after the shared opening. The first 5 characters will look right whatever happened, so they are not the part to check. The 51 characters after them are where a difference would show.

That habit will not catch everything, and the panel on what a check cannot reach is honest about the limits. It catches the case where you clicked a link that was not the link you thought, which is the case people never check for.

People arriving from ordinary browsing often want to make the experience faster or more familiar. Additions and adjustments generally cost more than they return, because they make one browser distinguishable from another and the whole value of using a common configuration is that it is common.

The board is not going to give a configuration walkthrough, and there are no screenshots here of any browser. The general principle is enough: leave it as it arrived, resist the urge to improve it, and treat anything that asks you to change a setting to make a page work with the same suspicion you would give any other unexpected instruction.

Every tile on this panel counts characters. That is not an accident or a shortage of imagination: the browser is not something this site can measure. No timing is collected, no request is made, and nothing here observes a load.

A panel about the browser that showed load times would be a panel showing invented numbers. So this one shows the strings involved, describes behaviour in words, and leaves the tiles honest.

Why does the same address sometimes load fast and sometimes not?

The path is rebuilt regularly and different paths perform differently. The address is identical either way, which is why speed says nothing about the address.

Does the browser remember which address I used?

Only in the ordinary way that browsers remember addresses. It has no concept of a set of related addresses and will not compare one against another.