From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Sat,  7 Aug 2004 01:56:43 +1000
From: George Michaelson <ggm@apnic.net>
To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu>
Subject: Re: [9fans] rc mystery
Message-Id: <20040807015643.373bb01f@garlic.apnic.net>
In-Reply-To: <285f606db6bcbbfc31f0a9f8cdf04593@9netics.com>
References: <20040806135100.5333c6df@garlic.apnic.net>
	<285f606db6bcbbfc31f0a9f8cdf04593@9netics.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Topicbox-Message-UUID: d1833be4-eacd-11e9-9e20-41e7f4b1d025

>>
>> But, don't you think there is something equally deceitful in setting trick cyclist
>> tests on people?
>>
>
>What's deceitful about asking somebody what does printf("%c\n",0["unix"]);
>print and why?  The "trick" is only a conversation piece.

Conversation pieces are best left for the dinner table, where wit, and verbal
fencing are appropriate. Personally, I find interviews a very degrading process
from both sides of the bench.

Its a very odd way to get one char out. You've made the compiler reduce a
string literal down to one char, to stuff into a stack, only to throw away
the rest of the string. I'm struggling to see why this is better than an
embedded /* we need to emit 'u' at this point */ comment which is probably less
work for the compiler, although I suppose it has to find the end of comment marker
rather than walk the embedded string, find its first element, and literal it.

Is it really efficient to array de-ref rather than embed the ASCII value directly?

I'm not convinced we wouldn't do better with a lottery. I am not a
psychologist and I cannot predict how people will respond to questions
where the 'desired' answer is not clear. It only gets worse if English is
not the first language. Maybe you don't hire non-nationals. Or do you 'test' them
in their own language? (perhaps C is, in these cases, the best choice..)

>
>Having been on the receiving end of it too, I sympathize; but I think
>it is a fair way to weed out some of the posers (with a degree or not).  I
>agree with what geoff said.  Code samples of previous work are probably
>more representative of one's true abilities.

Yes, I agree with that.

Faced with 40 or more shortlisted candidates for a post, its tempting to
find short paths to triage. But its also demeaning. I do it too.

-George

(to be less OT, gcc did at least make the same assembler for printf "%c\n", 117)