A list is a copy, not the thing

Two things share the word

There is the arrangement the market has made and considers current. Then there is the text on a page somewhere that reports it. Both get called the mirror list and only the first has any authority.

The second is a copy, and copies have properties: they were made at a time, by somebody, from a source, and they do not update themselves.

Why this framing is more useful than trust

The usual question about a list is whether the site is trustworthy. That question is nearly unanswerable and it produces long arguments that go nowhere.

Asking what kind of copy this is leads somewhere. Where did it come from, when was it made, has anything been added to it, and does it claim more than a copy could claim. All four are approachable.

What a copy can be right about

• The characters. A careful copy has exactly the strings its source had.
• Its own provenance, if it says where it came from.
• Its own limits, if it says what it did not do.

What a copy cannot be right about

• Whether the set has changed since the copy was made.
• Whether the source was itself current.
• Whether any string in it belongs to the market, as opposed to having been in the source.
• Anything about the state of any address, which no copy can carry.

This site is a copy too

The three strings printed at the top of this page were supplied and are published unchanged. This reference did not assemble them, does not open them, and would not know if the underlying arrangement changed.

Saying so is not a disclaimer bolted on at the end. It is the same fact that every study in this section is about, applied to the page you are reading.

The chain and where you sit on it

The arrangement

Held by the market. Changes without notice. Nobody outside can see it directly.

A published list

A copy of what its author had. As current as its source was at the moment it was made.

Your own store

A copy of a copy. The one you will actually use, and the one nobody else will ever refresh.

Why the third row is the important one

Your own copy is the one that gets used under pressure, and it is furthest from the source. Every problem in this section arrives at that row eventually, which is why the list you keep is the practical end of the section.

The single question worth asking of any list

Not is this site good, but what would this list look like if it were wrong. If the answer is exactly the same, which it usually is, then the list is not carrying the assurance you were hoping to read from it, and you should treat it as one copy among others.

Reading a page as an object rather than a source

Once you treat a list as a copy, the questions change shape. Not is this good, but what is this. Who made it, from what, how long ago, and what has been added to the strings that was not in them.

Those questions have partial answers, which is more than the trust question ever offers. A page that says where it got its strings has told you something you can weigh. A page that says its strings are the best has told you only that it is willing to say so.

It also changes what you do with disagreement. Two copies differing is what copies do, and the interesting part is never which one to believe but whether either contains something the other has never seen.

What this study is not. It is not an assessment of any site that publishes addresses, and no such assessment appears anywhere here.