Ask HN: How do you decide a name for your product or side project?

Assuming I can't just straight up think of and find an available .com in the exact name I'm after -

Take the subject - e.g. if I was starting a video hosting site I'd go with "vid", "video", "tube"

Chuck it in Impossibility [1] and see what combos bring out the best sounding results, repeat the same with Namemesh [2] and Namegrep [3]

Jot down all the viable options, sleep on it, register the one that still sounds appealing next day

[1] http://impossibility.org/ [2] https://www.namemesh.com/ [3] https://namegrep.com/

Full disclosure of course - I've not got any successful side projects out there yet. While I doubt it's the naming methodology.. it could be :)

• availability of related URL

• availability of a short name on twitter

• the name should describe the benefit of your product, not aspects of how it works. I made this mistake with my startup, which was originally named "Read The Rainbow" because it uses color gradients to make reading on-screen easier. But to someone who has never heard of our technology, hearing "rainbow" doesn't evince speed. We changed the name to "BeeLine Reader" to fix this and have gotten mostly positive feedback.

I'm founder of Tesults (at tesults.com) and the name is a portmanteau of test results. I put naming off until quite late and used a code name internally when developing the product. I hated the Tesults name on launch because I would have preferred to have a real English word. Having a newly coined word needs explaining and adds an extra hurdle. However there was nothing better I or anyone else could think of and we wanted to launch. After almost two years, I'm happy with it, it's short and easy to remember, especially for anyone who remember it is short for test results.

There are problems with almost any name though. Search for tesults on google and it will show you results for results instead and you have to click the link that says 'did you mean tesults?' This changed briefly and then reverted, I guess we are not big enough yet and google found people really did mean results and made a typo :-) This is not a huge problem though because new users don't find us this way.

We have some plans in the pipeline to do things adjacent to test results reporting and the name Tesults will be basically meaningless in that context but really so are so many names when you think about them. I don't think the name is that important, I do like having the .com though as the Paul Graham essay suggests, however even that is not a requirement for success as others have proved.

By setting aside the availability, it should be rememberable (for me it is either an animal or an object/person that represents the product/project). Lately, I chose the name Sandglass for a message queue (https://github.com/celrenheit/sandglass). The reason is that this project depends on an ID generator that I named sandflake (in reference to twitter's snowflake). So I chose the name sandglass because it is kind of like a glass with a lot of sandflakes flowing through it (like a message queue).

Another name I used was Lion for an http router but for no particular reasons :)

My projects so far are open source, and have taken two approaches for naming:

Either reuse a name of different context that kind of maps (eg a flexible, versatile framework/library called Bamboo)

Or literally just what it does: recently I’ve been working on projects called: Koalephant {Failover,Certificate,Proxy,PXC} Manager

Koalephant is the name of my company, the tools are management/sync helpers to setup failover, tls certificates, tcp and http proxying, and percona Xtradb cluster.

Related (a bit) http://www.paulgraham.com/name.html

I normally find 2 discriptive words, concatenate them and repeat until I find an available .com (e.g. wishminer, for a tool that mines wishes).

The name should be easy to pronounce and spell. Often repeated vocals make the name more memorable. You can test the name by mentioning it to people before describing the app and ask what they'd expect. Then check a day later which name they remembered.

I also suggest checking for trademarks as soon as you come up with something. It's really disappointing to let a name "grow on you" only to realise it's taken when you're ready to launch..

I use the global brand database: http://www.wipo.int/branddb/en/

As far as websites go, this is getting harder all the time. It seems the majority of domain names are owned by domain squatters.

We create external camera pods for aircraft, which themselves can't fly. So we name them after flightless birds.

Tomorrow I'll be flying 'Takahe':

https://en.wikipedia.org/wiki/South_Island_takah%C4%93

This has always been a huge problem for me before I've came to the idea of using Markov chains to generate new words :-) There are many ready-to-use implementations available. Just import a dictionary/text, generate a list of new words and choose the one you like.

All my projects have two names, internal code name usually a god or star name and external name.

So for example our ERP at work is internally Minerva and it's upgraded version is arcturus.

Business side it's just referred to as wholesale.

Sometimes, I translate the English word/phrase describing the product into a foreign language. I usually try a bunch of languages (Latin is generally excellent) before deciding on what I like best.

Smash together random letters until it sounds nice.