Generators — QR code, password and MD5 hash
Make a QR code for any link, a password you can actually use, or an MD5 hash of a string.
A generator produces output from a rule instead of from a stored list. Give it a specification and it builds items that match. This page covers what generators are good for, how they fail, and how to check output you did not write yourself.
Tools in this section
Other sections
How to use Generator
Every generator needs three things decided before it runs: what shape the output takes, how many items you need, and how you will tell good output from bad. Skip the third and you will not notice when the generator is producing something subtly wrong.
Define the format precisely — length, allowed characters, required structure. Decide whether duplicates are acceptable, because most generators allow them by default. Generate more than you need if you are going to filter. Inspect a sample by hand before trusting the batch.
What generators are actually for
The common uses split into a few groups. Filling — placeholder text, sample records, test data that needs to exist but not to mean anything. Enumerating — producing every combination that fits a rule, where writing them out by hand would be error-prone. Securing — passwords and identifiers that need to be unpredictable. Varying — creating many versions of something from a template.
These have different requirements. Test data can be repetitive and it does not matter. A password that is repetitive is broken. Knowing which category you are in tells you what to check. A generator will happily produce ten thousand items that are all wrong in the same way. The volume is what stops you noticing.
The uniqueness problem
Generated items collide more often than people expect. If a generator draws from a limited set of possibilities, duplicates start appearing far sooner than the size of that set suggests — the chance of a collision grows with the number of pairs, not the number of items. A few hundred items from a set of thousands will usually contain a repeat.
If items must be unique, check for duplicates rather than assuming. Enlarge the possibility space rather than adding retries — retries hide the problem without solving it. For identifiers that must never collide, use a format designed for it rather than a short random string.
Checking output you did not write
Generated content needs a different kind of review than written content, because the errors are systematic rather than scattered. One mistake in the rule appears in every item. This makes it easier to find if you look for it and easier to miss if you spot-check casually, because every sample confirms the same pattern.
The useful approach is to check the extremes rather than the middle: the first item, the last item, the shortest, the longest, anything at a boundary. Errors in a rule show up at edges. A hundred items from the middle of the range will all look fine even when the rule is wrong.
Where generators stop being appropriate
A generator is the wrong tool when the output needs to be meaningful rather than merely valid. Placeholder text that ships to real users, sample data that gets mistaken for real records, and content generated to fill space rather than to say something all cause more trouble than the effort they saved.
The failure mode is specific: generated filler is designed to look plausible, so it passes casual inspection indefinitely. It gets discovered when someone tries to use it for something, which is usually much later and much more expensive than catching it at the start. If output is going anywhere a person will act on, it needs a human decision behind it — not just a rule that produced something well-formed.
Questions and answers
- Will a generator ever produce the same result twice?
- Yes, unless it is specifically tracking what it has already produced. Most generators draw independently each time, so repeats are expected rather than a fault.
- How long should a generated password be?
- Longer than feels necessary, and drawn from a cryptographic generator rather than an ordinary one. Length contributes more than character variety, and the generator's quality matters more than either.
- Can I use generated placeholder text in a live page?
- You can, and it will be found — usually by a visitor rather than by you. Placeholder text is for layout during development, not for publication.