It's worth adding that you should reach out to the author first if you find 
such a package in npm. There's a good chance that they're not "squatting" 
the name intentionally - they probably just started a project excitedly and 
didn't push as fast as they'd hoped.

I had a "dynamo" and "task" in npm without little or no code. I was 
contacted by jed who had a dynamo client ready to go and tkellen who wanted 
task for a gruntjs related project. - in both cases I gave up the names.

No drama, no need call anyone a squatter, no need to get Isaac involved, 
good times.


On Saturday, December 15, 2012 11:10:23 AM UTC+11, Isaac Schlueter wrote:
>
> TL;DR - Package squatting is not allowed.  If you sit on a package 
> name and don't publish code, it'll be deleted without warning. 
>
> ---- 
>
> npm package names have always been a "first come, first served" 
> system.  I think this is generally good.  It incentivizes early 
> adopters, which is important, because they're the most valuable users. 
>
> It also is a powerful motivator to *actually write code*. 
> Traditionally, the pattern has been that if you think of a great name, 
> well, you'd better ship something, then!  We never had to have an 
> explicit rule about publishing empty packages, for the same reason 
> that you don't have to have an explicit rule about leaving garbage on 
> the floor in someone's house.  It's just not something you'd do. 
>
> It is hard to really comprehend *just how many* possible package names 
> there are.  The math isn't that hard, but the actual numbers are 
> mind-boggling.  Even just limiting to 4 characters, there's 
> 36*38*38*38 potential names.  That's about 100x the number of npm 
> packages that have ever been published.  And yes, it includes stuff 
> like 0-_-, but still.  You don't have to limit it to 4 characters.  At 
> 8 characters, there's 4118960973312 potential names.  If you got TJ 
> Holowaychunk and Raynos and Substack and architectd and Dominic Tarr 
> together and managed to get them to publish a package with a unique 
> 8-character name every MILLISECOND, it'd take OVER 130 YEARS to ever 
> use them all up!  Absurdity aside, there are millions and millions of 
> common words and clever spellings.  We'll never run out.  There is no 
> scarcity. 
>
> (To the alpha-nerds in the audience: this is not a challenge.  Please 
> don't DOS the registry. ;) 
>
> The official policy for handing name disputes mostly focuses on 
> abandoned modules, confusing name collisions, and other cases where 
> two parties both actually have published (or want to publish) working 
> code.  There's never been a need to make the "no seat saving" 
> explicit, because it just didn't happen much.  Disputes between two 
> authors have generally always been handled pretty easily.  The 
> occasional "no code here" module was always a mistake or an oversight, 
> and promptly cleaned up. 
>
> Lately, probably owing to the increase in our community's size, or the 
> increase in npm's popularity, or some combination of factors, I've 
> been seeing a lot more cases where someone asks to use a module, 
> pointing out that the author isn't using it, and they're told, "No, 
> that's something I'll be publishing at some time in the future."  When 
> I've stepped in and made a ruling, pointing out that **npm is for node 
> modules that exist, not for node modules that don't exist**, the 
> squatters have in some cases reacted with surprise and frustration. 
> It's gotten un-amicable. 
>
> If you think I'm talking about you, you're probably right.  But I 
> wouldn't bother to write this if it had been an isolated incident, so 
> you're not alone.  I don't want to single anyone out, and it's 
> happened enough times that clearly there's some widespread confusion 
> about what's ok and what isn't.  You're not a bad person.  You didn't 
> know. 
>
> To be fair, yes, this was never a part of any "official" policy in an 
> explicit way.  So, as of 
>
> https://github.com/isaacs/npm/commit/993abe412db791f9460932371b39642817523aab 
> (and this email) it's explicit.  (The great thing about npm being a 
> dictatorship is that it doesn't ultimately matter what any policy 
> says.  My interest is for the node community to write code, so 
> anything that stands in the way of that gets cut down, policy be 
> damned.) 
>
> It is very easy to rename a module.  It's a matter of changing *a 
> single field* in a JSON doc.  If it takes a few months to build 
> something, and someone publishes something in with the same name in 
> the meantime, that means a few things are likely: 
>
> 1. You took took long.  Go faster.  Which probably means: 
>   1a. You tried to write something that was too big.  Build smaller. 
> 2. The name you chose is too vague and doesn't tell much about your 
> problem domain.  Which probably means: 
>   2a. Your problem domain is ill defined.  (Is it an MVC framework? 
> CMS platform? Stop right there.) 
>
> If that happens, just pick a new name!  There are literally zillions 
> to choose from.  It doesn't matter how long you've slaved over the 
> code to make your module, with the hope of using that name.  They got 
> it over the finish line first.  That's how "first come first served" 
> works. 
>
> This is the warning.  If you're reading this, and you have something 
> that you're "intending" to publish "soon", hurry up!  If your module 
> is just a placeholder, it'll be removed the first time someone calls 
> attention to it. 
>
> It pains me to have to do administration of any sort.  Really.  You 
> have no idea how lazy I am when it comes to petty things like this, 
> and I don't enjoy being a hardass.  But npm is not for owning words. 
> It's for sharing code.  Hack or get out of the way. 
>

-- 
Job Board: http://jobs.nodejs.org/
Posting guidelines: 
https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Reply via email to