2009/200 Solaris Simlinks

Scott Rotondo Scott.Rotondo at Sun.COM
Fri Mar 27 10:06:32 PDT 2009


James Carlson wrote:
> Scott Rotondo writes:
>> Rishi Srivatsavai wrote:
>>> I don't think it is used much elsewhere (our project gate is the 5th result
>>> in Google) and I doubt it would cause much confusion given that this
>>> is a testing resource. But if it does we can certainly re-visit the 
>>> choice of name in future.
>> In that case, I think you'd be doing your future users a favor if you 
>> pick a different name: emlinks, pseudolinks, simulinks, ... whatever.
> 
> I think y'all have seriously lost your sense of humor.  :-/
> 
> I happen to think simlink is a good name, as it's *clear* from the
> context that this can't possibly be the same thing as "ln -s" (I can't
> see how anyone could be confused), and it's both memorable and
> evocative.

If you know the context, it's clear. My day-to-day experience is that I 
frequently have conversations where not everyone has perfect context. 
I'm imagining this kind of exchange:

A: I was testing NFS over [insert arcane networking feature]. I copied a 
few GB just fine, and then I hit an assertion failure.

B: Were you using simlinks?

A: No, just regular files. Why would that matter?

B: I mean the network interface.

A: No, it's a regular device node. [pause] Wait a minute, you're right. 
The /dev entry is a symlink to something in /devices. I never noticed 
that before. What do I do about that?

... and so on until enlightenment occurs.

> 
> But, frankly, I don't think the argument is worth a tinker's (or
> hacker's) damn.  If you absolutely insist that it's unacceptable with
> the name chosen by the project team, then please provide one that
> isn't either ungainly to use (pseudolinks) or burdened with a tin ear
> (emlinks).  (Jury's out on simulinks ... it makes me think that Major
> League Baseball is going to prohibit use without their express written
> consent.)

I didn't absolutely insist, I just made a suggestion.

Personally, I think simulinks and emulinks are both fine replacements.

	Scott

-- 
Scott Rotondo
Principal Engineer, Solaris Security Technologies
President, Trusted Computing Group
Phone/FAX: +1 408 850 3655 (Internal x68278)



More information about the opensolaris-arc mailing list